IRC logs for #openttd on OFTC at 2024-07-03
⏴ go to previous day
00:42:54 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
00:49:19 *** k-man has quit IRC (Quit: WeeChat 4.3.3)
01:05:49 *** Wormnest has joined #openttd
01:29:01 *** Wormnest has quit IRC (Quit: Leaving)
02:04:01 *** herms6 has quit IRC (Quit: bye)
02:43:04 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
02:49:31 *** godbed_ has joined #openttd
02:52:51 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
02:52:56 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:00:01 <talltyler> andriydohniak: There are a few valid suggestions in that comment but if the tone bothers you (it bothers me) then you are not obligated to read or respond to any of it. 🙂
03:00:01 <talltyler> I will take a look tomorrow at parsing the hyperbole into actionable suggestions for you.
03:12:49 *** reldred has joined #openttd
03:13:03 <reldred> Surprised they haven’t been blocked yet
03:14:32 <goddess_ishtar> is Zorg's MO just "this is cool, but you should totally fix this entirely different thing instead"
03:16:48 <reldred> Best ignored and move on
04:42:44 <DorpsGek> - Update: Translations from eints (by translators)
05:33:20 <johnfranklin> Zorg is saying "I am sorry for may English. I know is bed." But Zorg can write such long essays very often.
06:28:19 *** alfagamma0007 has joined #openttd
06:44:35 <pickpacket> Zorg is a cooler name
06:51:14 <goddess_ishtar> when he comments on another PR, is that a Zorg rush?
06:52:21 <truebrain> No, a "dddooouubbbbllleeeee zorg"
06:55:44 <goddess_ishtar> you're right, a Zorg rush would be when he comments on a PR with several paragraphs of text immediately after it is posted
06:58:43 <truebrain> `waiting for updater: waiting for container: context deadline exceeded` <- it took dependabot 30 minutes to figure out something went wrong, lol
07:00:59 <truebrain> another job here just hang for 6 hours
07:01:08 <truebrain> GitHub was not having a good day, yesterday
07:05:39 <truebrain> for the next hour or so, I will be redeploying some new versions on our backends. Expect some minor downtime on services like bananas, wiki, ... (few minutes, at most)
07:11:30 <reldred> I accidentally put an hour long dent in our Zabbix logs, tried to do performance optimising, didn’t realise by increasing all the worker counts I’d hit the db max connections limit…
07:12:11 <truebrain> unexpected downstream effects 🙂
07:12:14 <truebrain> only an hour is not bad, in that case
07:12:44 *** DorpsGek has joined #openttd
07:12:44 *** ChanServ sets mode: +o DorpsGek
07:13:55 <reldred> Forty minutes technically,
07:14:35 <reldred> Doesn’t help there was an exec who wanted some numbers/graphs, took the previous hour instead from grafana
07:15:58 <truebrain> dependabot is doing an aweful job lately
07:16:10 <truebrain> timeouts, not updating, saying it is going to recreate things, but actually not
07:21:24 <truebrain> especially as I can't really tell if it has all the latest dependency updates now
07:21:29 <truebrain> any updates better than none
07:28:29 *** HerzogDeXtEr has joined #openttd
07:29:32 <truebrain> okay, everything seems to have survived
07:29:49 <truebrain> now I have to look at some AWS infra parts .. as I got a stern email from AWS that I didn't do a thing they told me to do
07:30:27 <truebrain> `You were previously contacted in May 2024 regarding Amazon Elastic Container Service (ECS) and the introduction of tagging authorization for resource creation. Your account is continuing making tag-on-create calls without proper ecs:TagResource permissions since April 1, 2024.`
07:30:32 <truebrain> and I was totally unaware we still used ECS 😛
07:46:58 <andriydohniak> reldred: He is right on a lot of points though.... and I am not happy about it
07:47:35 <andriydohniak> Yea, his suggestion to fix the small fonts everywhere is just ridiculous, but alsmost everything else, he is kinda right
07:48:52 <andriydohniak> and now I am having an existencial crisis about this pr
07:52:31 <andriydohniak> At least I will make the minimum number of rows bigger, so the graph is taller, and change the buttons to be vertically stacked. I might try making a fork of my pr and experimenting with embeding the widget in the graph windows... again..., and the search bar on the companies, it is probably gona stay, it doesn't take up much space, and for the sake of consistency it is fine, and sometimes useful
07:52:31 <andriydohniak> even for 15 companies
07:54:49 <andythenorth> hmm nile might do grf translations? 😮
07:55:20 <truebrain> andriydohniak: with Zorg it rarely is about whether they is right or not. It has much more to do: is it in scope of this PR, and is it even reasonable to ask. Often they asks for complete rewrites of the whole game to make it happen
07:55:29 <truebrain> sometimes it is better to make a small step, which might be slightly in the wrong direction
07:55:37 <truebrain> than to spend days, months, years doing the "absolutely perfect thing"
07:56:07 <andythenorth> it's poor feedback, and the occasional nugget of gold in the coal dust is rare rare rare
07:56:13 <andythenorth> and not worth the effort
07:56:33 <truebrain> so just be careful to not make the PR totally unreviewable, just because "they have a point" 🙂
07:57:29 <andriydohniak> truebrain: Yea, I probably will just tweak some values like described in my comment above, which will adress some of the points he had, but I won't fully rewrite it ... again
07:58:54 <andriydohniak> so I will make the legend taller and thinner, and will take a look at how the minimal size was determined in the old legend, and if applicable will try to implement that
08:06:34 <andriydohniak> And yes his tone is absolute shit, and this was personal:
08:06:34 <andriydohniak> > > JustLinuxUser: 2. It is impossible to make the side widget expand untill something, it either expands infinitely or not at all, and the max possilbe cargo name length is just too wide
08:06:34 <andriydohniak> > Don't cheat reality. ;) In a way that is impossible for you, the current payment rate graph window has been working for almost 30 years (?) The width of the list adjusts to the longest name, and then when you stretch the window, only the size of the part with the graph changes.
08:06:34 <andriydohniak> But I still think that his oppinion is important to consider, because it is very much resonable
08:07:41 <truebrain> okay, next up I am going to rebuild our VMs running in the backend. This causes minor interruptions throughout our backend services, like disconnecting TURN sessions.
08:09:57 <andriydohniak> truebrain: Are you using some kind of provisioning script like Ansible / OpenTofu?
08:10:10 <truebrain> Pulumi and AWS AutoScaling Groups
08:14:58 *** DorpsGek has joined #openttd
08:14:58 *** ChanServ sets mode: +o DorpsGek
08:17:27 *** audigex has quit IRC (Remote host closed the connection)
08:17:27 *** wensimehrp has quit IRC (Remote host closed the connection)
08:17:27 *** DorpsGek_vi has quit IRC (Remote host closed the connection)
08:17:27 *** truebrain has quit IRC (Remote host closed the connection)
08:17:27 *** reldred has quit IRC (Remote host closed the connection)
08:17:27 *** alfagamma0007 has quit IRC (Remote host closed the connection)
08:17:27 *** johnfranklin has quit IRC (Remote host closed the connection)
08:17:27 *** _jgr_ has quit IRC (Remote host closed the connection)
08:17:27 *** _glx_ has quit IRC (Remote host closed the connection)
08:17:27 *** talltyler has quit IRC (Remote host closed the connection)
08:17:27 *** merni has quit IRC (Remote host closed the connection)
08:17:27 *** silent_tempest has quit IRC (Remote host closed the connection)
08:17:27 *** locosage has quit IRC (Remote host closed the connection)
08:17:27 *** andriydohniak has quit IRC (Remote host closed the connection)
08:17:27 *** peter1138 has quit IRC (Remote host closed the connection)
08:17:27 *** goddess_ishtar has quit IRC (Remote host closed the connection)
08:17:27 *** kuhnovic has quit IRC (Remote host closed the connection)
08:17:27 *** andythenorth has quit IRC (Remote host closed the connection)
08:17:37 *** DorpsGek_vi has joined #openttd
08:18:02 *** truebrain has joined #openttd
08:18:02 <truebrain> right .. that should be most of it
08:18:25 <truebrain> live-migration for containers still isn't a thing, sadly 😛
08:19:16 <dwfreed> don't think criu ever got very far
08:20:29 <truebrain> and cutting the last few connections forcefully ... sorry TURN users 😦
08:21:29 <truebrain> still the best way to do OS updates btw; just spin up a new VM 🙂
08:21:52 <truebrain> either it gets online and everything is fine. Or it doesn't, and nothing breaks 😛
08:23:25 <truebrain> and Nomad finished its new leader selection \o/
08:27:13 <truebrain> a piece of software
08:34:06 <pickpacket> Ah, software. Such magical things
08:37:28 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
08:41:59 <peter1139> I wanted to do dockable windows for the new picker stuff.
08:42:36 <peter1139> That's how I ended up with the concept of a base window that can be reused in different ways with different parts visible.
08:43:38 <peter1139> Ideally it would've been a widget that handles event as needed, but widgets don't handle all events, and data passing becomes a nightmare.
08:44:26 <peter1139> Oh, context is this graph key PR.
08:50:09 <truebrain> owh, and just a kind reminder for anyone with an open SSH port to the outside world: maybe update your OpenSSH 🙂
08:54:58 *** reldred has joined #openttd
09:00:19 *** andriydohniak has joined #openttd
09:00:19 <andriydohniak> peter1139: I decided to make it a separate window, because making both the company legend and the cargo legend being an integrated widget would be very hard to make work well, because:
09:00:19 <andriydohniak> 1. I didn't know it was even possible to detect the max size of string to adjust the size of the buttons
09:00:19 <andriydohniak> 2. In case of companies, the company name limit is just too long, so even if I did do that, if somebody would have changed the name of the company to a long name it would make the graph widget unusable for everybody
09:00:19 <andriydohniak> 3. I didn't want to do 1 thing for companies and another for the legend
09:05:14 *** ialokin has quit IRC (Ping timeout: 480 seconds)
09:07:30 <peter1139> I wasn't commenting on any specifics, just that I investigated ways for allowing for both built-in and separate widget comoponents, and letting the player decide.
09:24:25 <peter1139> Me: "Here's the specification of a simple CSV format to send us."
09:24:39 <peter1139> Them: "Here's a file that almost but not quite matches your specification."
09:25:58 <peter1139> Trailing commas on some of the lines.
09:26:21 <truebrain> nothing you can't fix for them, I am sure!
09:26:52 <peter1139> I check that the right number of fields are present because I don't want to be misprocessing data.
09:27:46 <peter1139> Ah well, yes, I now check if there's a trailing comma, i.e. an extra field is permitted as long as it's blank.
09:27:59 <peter1139> And yes, I know, be lenient in what you accept, and strict in what you send.
09:28:06 <peter1139> But there are limits :)
09:29:30 <truebrain> just bill them an hour or so for "being stupid" 😛
09:43:25 <merni> peter1139: Perhaps they used Excel to create the file :)
09:43:37 <peter1139> truebrain, well, this is also the same one use prices with 4 decimal places...
09:45:36 <peter1139> Also sent a file where 7.74 + 33.35 = 41.08 ...
09:47:53 <peter1139> Maybe they used an LLM to calculate it.
09:49:06 <dwfreed> especially if using "banker's rounding" (round to even on split)
09:49:28 <dwfreed> ie, if the math came to 41.085, you'd round to 41.08
09:50:37 <peter1139> @calc 6 * 1.2895 + 24 * 1.3894
09:52:38 <peter1139> I was well aware. It's not just rounding errors. It's errors due to inappropriate rounding. :-)
09:52:58 <dwfreed> each half of that addition rounds up, but taken together you get the round down
09:53:56 <dwfreed> but yeah, intermediate rounding quickly bites you in the ass
09:55:42 <peter1139> In this context, *not* rounding at each stage is wrong.
09:56:06 <peter1139> Why do I never have enough USB ports...
09:56:37 <dwfreed> I once was really annoyed with lack of USB A charging ports, so I bought a 10 port anker charger
09:58:00 <merni> the teenager in me wants to add a letter
09:58:41 <peter1139> I have 4 upward facing USB ports on my desktop PC case.
09:59:09 <peter1139> They are of course unrealiable due to the inevitable build up of dust.
10:00:31 <peter1139> As if I don't have multiple hubs already :D
10:11:42 <peter1139> Hmm, perhaps if I packed away some of my synthesizers...
10:12:20 *** johnfranklin has joined #openttd
10:12:21 <johnfranklin> Can you see pictures in IRC?
10:13:04 <peter1139> I can open links to pictures.
10:20:57 *** KK9 has quit IRC (Remote host closed the connection)
10:27:17 <andriydohniak> peter1139: It might be dust build up, or if you have a cheap/old pc case it might be the cable that connects the front pannel ports to the mother board
11:10:00 <peter1139> Is a hotdog a meal or a snack?
11:24:58 <truebrain> depends on when it is consumed
11:58:06 <andriydohniak> peter1139: both
12:01:11 <andriydohniak> I have an idea, if you eat it of a plate, it's a meal, if you hold your food, it's a snack
12:03:48 <LordAro> what about a plate of hot dogs?
12:12:13 <truebrain> You can't make a plate out of hot dogs .. at least .. not while they are also still eatable, I think
12:12:22 <truebrain> if you put them in epoxy you can ofc
12:12:26 <truebrain> but that seems defeating the purpose
13:09:52 <andriydohniak> LordAro: but hot dogs on a plate, depends. If you eat them from the plate directly with a fork, it's a meal, if you take it and eat from your hands it's a snack
13:10:44 <andriydohniak> truebrain: I am shocked that when you see "plate of hot dogs" you think of a plate made with hot dogs 🤣 🤣
14:14:37 *** audigex has joined #openttd
14:14:37 <audigex> peter1139: NZXT H700i?
14:14:45 *** andythenorth has joined #openttd
15:00:13 <peter1139> There's a PR to fix that.
15:00:31 *** Wormnest has joined #openttd
15:08:39 *** HerzogDeXtEr has joined #openttd
15:09:27 <peter1139> Still too many bloody cars in this bike race.
15:24:11 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
15:25:49 <pickpacket> Hmm. Since _glx_ fixed the compiler warnings I guess my next goal for the code is filters in the town listing
15:27:09 <pickpacket> After the two gamescripts and a bunch of non-OpenTTD stuff on my list...
15:37:49 *** gelignite has joined #openttd
16:16:11 <peter1139> So this client with the CSV file went live.
16:16:27 <peter1139> And they sent a completely different XML format instead...
16:28:28 <truebrain> at least they are consistent 😛
16:30:04 <truebrain> (being consistently inconsistent with the specs is also consistent!)
16:30:53 *** Smedles has joined #openttd
16:31:52 <andriydohniak> peter1139: I think libreoffice can import CSV and xml, and you can export in whatever format you like, but probably not json
16:32:39 <peter1139> But no, that's 100% irrelevant.
16:49:25 <andriydohniak> *Fixed all the issues, (and thanks for the wonderfull suggestion to use ScrollTowards), mentioned by Peter, also made the buttons be on top of each other, and now the graph is thinner and longer, also only opens automatically for the cargo graphs
16:49:54 <truebrain> is that again a challenge?! 😛
16:51:46 <andriydohniak> truebrain: I bet it doesn't matter what it is, you ain't gona brake it anyway!
16:52:28 <truebrain> oeh, I see a frosch123 . Hi frosch123 . I have nothing else to say, mind you. Just: hi 🙂
16:52:55 <andriydohniak> This is how it looks now btw when you open the cargo graph when no windows are open
16:53:20 <truebrain> now the graph window should be ever so slightly higher by default 😛 This hurts my OCD! 😄
16:53:29 <truebrain> (I am joking, to be clear)
16:53:49 <andriydohniak> It is clear, and allign from the top is supirior anyway
16:54:04 <andriydohniak> left -> right, top -> bottom, my western brain works like that
16:55:25 <andriydohniak> But also, the cargo graph is smaller then all other flags, should I change that?
16:55:52 <truebrain> if that was already the case, absolutely not in this PR
16:56:09 <truebrain> always minimize the work done in a single PR
16:56:22 <truebrain> we rather have 20 small PRs, than one "fixes everything" PR 🙂
16:56:26 <andriydohniak> It's too late, but for the future, noted
16:57:16 *** silent_tempest has joined #openttd
16:57:16 <silent_tempest> It's too late? What do you mean?
16:57:53 <andriydohniak> silent_tempest: I am joking, but this PR is +831 -341 loc
16:58:25 <silent_tempest> Line count is not a very meaningful metric
16:58:49 <andriydohniak> Yea, and I tried to stay on point, so this PR is pretty focused
16:58:57 <truebrain> andriydohniak: no worries; just to give a bit of context: we sometimes get large PRs, that have like 60% good stuff in it, and 40% less-good stuff. It is hard to figure it out from there. If it would be 10 PRs, as weird example, it would be easier to accept the 60%, and not accept the 40% 🙂 So changing the width of a window is such example. That can be reviewed much easier, on its own, than it
16:58:57 <truebrain> would in a massive PR 🙂
16:59:59 <peter1139> How did I ever get WidgetDimensions in...?
17:00:19 <truebrain> brute force, I am sure 😛
17:01:24 <peter1139> Used a sockpuppet account...
17:01:44 <peter1139> Or we all just got fed up of it looking like shit before :D
17:01:59 <truebrain> `sudo git push -f`, right?
17:03:15 <andriydohniak> `sudo ssh root@github.com git clone <your repo> /tmp/my_repo && mv /repos/OpenTTD/OpenTTD`
17:04:09 <andriydohniak> s in ssh stands for "Secure", right? right?
17:04:21 <andriydohniak> He probably has a 0 day patch 🤣
17:06:08 <truebrain> you mean the one that got released a few days ago?
17:06:11 <truebrain> now it all makes sense
17:12:31 <andriydohniak> truebrain: He could't keep it a secret, he is a chatty one
17:36:09 *** talltyler has joined #openttd
17:36:09 <talltyler> andriydohniak: The `Hide all` button being wider than the `Show all` button due to the resize handle is a strange look. Not sure what should be done instead but I can’t unsee it. 😛
17:38:19 <truebrain> add a fake empty panel to the right?
17:39:05 <talltyler> Doesn’t Peter have a patch that makes scrollbars the same width as resize handles?
17:39:29 <talltyler> (Not that it would help here, but that’s the next thing we’ll notice 😛 )
17:39:32 <truebrain> the other day I found out peter doesn't have a patch for EVERYTHING
17:39:37 <truebrain> still not recovered
17:39:45 <peter1139> I probably have one but forgot about it.
17:40:01 <truebrain> btw, nice way of dealing with Zorg talltyler
18:02:47 <andriydohniak> talltyler: Would you prefer a bit wider window but them being side by side?
18:03:09 <andriydohniak> I saw some other window doing the same thing
18:06:59 <silent_tempest> So the emscripten build target is what the preview build is?
18:33:36 *** gelignite has quit IRC (Read error: Connection reset by peer)
18:33:51 *** gelignite has joined #openttd
19:49:41 *** godbed is now known as debdog
20:04:49 <talltyler> andriydohniak: I think so, yes
20:07:13 <andriydohniak> You got lucky, and in english, the title takes up more space then 2 buttons combined anyway, so it's 0 cost for english
20:08:11 <andriydohniak> but now the resize handle is a bit wider then the scrollbar, and that's not fixable, I don't think
20:09:47 <andriydohniak> It's hard being a perfectionist in the imperfect world 🤣
20:09:47 <talltyler> That is the case for all scrollbars and resize handles 🙂
20:10:27 <talltyler> (Feel free to do a follow-up PR that fixes that, now that you’re a GUI expert 😛 😛 )
20:11:20 <andriydohniak> talltyler: Yea, no I kinda get how it works now, but I am basically as far from an expert as a guy working construction in a 1800s
20:15:40 <andriydohniak> talltyler: The easy way to fix that would probably to just make the scrollbar a bit thicker and make it reuse the sizing code from the resizebox
20:16:58 *** gelignite has quit IRC (Quit: Stay safe!)
20:18:22 <talltyler> Thinking about consistency, I took a look for other windows which have enable and disable all buttons:
20:18:22 <talltyler> The smallmap has Enable all/Disable all buttons, horizontally.
20:18:22 <talltyler> Cargo flow legend has All / None, buttons, stacked.
20:18:22 <talltyler> Settings has Expand all / Collapse all, horizontally.
20:19:18 <talltyler> What if you did All / None, in that order, to match?
20:19:29 *** goddess_ishtar has joined #openttd
20:19:29 <goddess_ishtar> andriydohniak: can we have something like the key in this image for cargo link graphs please
20:20:08 <talltyler> Not in this PR! 😄 (but a follow-up would be lovely)
20:20:27 <andriydohniak> You guys just showed how OpenTTD sucks you in 🤣
20:20:43 <andriydohniak> You just do all the sucking for OpenTTD 🤣 🤣
20:21:01 <goddess_ishtar> the UI is in desperate need of improvement and standardisation
20:21:10 <goddess_ishtar> and I didn't notice until I actually tried using the graphs lmao
20:22:17 <andriydohniak> After this PR gets merged, I don't mind dedicating ~1h per day slowly working on the UI, but I am very much not a graphics designer, kinda the opposite
20:22:37 <andriydohniak> for me function > form, except for 1 pixel of cases 🤣
20:24:43 <andriydohniak> I need to save some time to actually go and play OpenTTD though
20:25:26 <andythenorth> form and function are indivisible 🙂
20:25:35 <andythenorth> design is how it works
20:26:32 *** nielsm has quit IRC (Ping timeout: 480 seconds)
20:26:58 *** locosage has joined #openttd
20:26:58 <locosage> I probably spent years by now on trying to improve ui
20:27:02 <locosage> well, more like ux than ui
20:27:59 <andriydohniak> I would like not to make life long commitments to OpenTTD yet 😄
20:32:40 <andriydohniak> truebrain: I am telling you someone is playing a Jia Tan rn
20:34:04 <_jgr_> Eh, this is really not an exotic bug you could hide anything interesting in 😛
20:34:06 <truebrain> _glx_: dependabot is really stuck on the `actions` repo. I hope it self-fixes over time, but it has been stuck for weeks now. Not sure what is going on there. There are updates .. it is just not working correctly
20:34:33 <andriydohniak> _jgr_: Just like Jia Tan adding boring cases to the test code nobody was looking at! 🤣
20:35:54 <truebrain> maybe we have to bump the dependencies that are "stuck" ourself .. maybe that unstucks it
20:36:03 <truebrain> really weird bug, with no clear reporting about what is breaking
20:39:26 <_glx_> there are some errors, but yeah
20:39:58 <_glx_> yup you don't have access to it 🙂
20:42:34 <truebrain> _glx_: Yeah, but they are weird errors .. timeouts ... very odd
20:43:39 <truebrain> Will do a manual update tomorrow ... it has one job ...
20:45:23 <_glx_> maybe monthly is too large for npm
20:52:09 <truebrain> Other repos of mine are fine
20:52:38 <truebrain> I had other PRs stuck btw; dependabot is just weird. Managed to fix the other cases. This remains stuck 😄
20:57:01 <_glx_> could be some weird recursion in the deps ?
20:57:35 <_glx_> or incompatible updates between them
21:30:57 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:43:32 <truebrain> Why does that description feel AI generated? 😛
21:43:44 *** ChanServ sets mode: +v tokai
21:44:38 <truebrain> And a nice quick response, that is nice 🙂
21:50:32 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
21:52:58 <peter1139> I do actually have a patch for the resize box issue somewhere.
21:53:11 <peter1139> The scrollbar width cannot be increased, that is sacred :-)
continue to next day ⏵
