IRC logs for #openttd on OFTC at 2021-08-31
⏴ go to previous day
01:18:31 *** Beer has quit IRC (Read error: Connection reset by peer)
01:39:11 *** Wormnest has quit IRC (Quit: Leaving)
02:02:25 *** _aD has quit IRC (Quit: leaving)
02:37:44 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:26:51 *** tokai|noir has joined #openttd
05:26:51 *** ChanServ sets mode: +v tokai|noir
05:33:40 *** tokai has quit IRC (Ping timeout: 480 seconds)
05:50:47 <Rubidium> Gustavo6046: AFAIR the speed of (un)loading is not linked to a track type, but rather to the individual vehicles. Similarly the faster (un)loading only has a secondary effect on the station rating, as there should be fewer people waiting (which influences that rating a lot), though having multiple slower loading trains running in parallel will also get you a lower rating
06:18:15 *** andythenorth has joined #openttd
06:44:36 *** sla_ro|master has joined #openttd
07:26:50 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
07:31:11 <andythenorth> seems GSList.Sort() can't sort cargo labels lexically
07:31:19 <andythenorth> sometimes works, sometimes doesn't
07:42:02 <andythenorth> I really hate GS currently
07:42:08 <andythenorth> if I log a cargo I get an int
07:42:25 <andythenorth> but if I pass the cargo to GSCargo.GetCargoLabel(cargo) to get the label
07:42:30 <andythenorth> I get an error that I've passed a string
07:49:22 <Xaroth> do you get an `1` or a `"1"` int ?
07:49:59 <Xaroth> `log(1)` vs `log("1")` gives the same output, but different types.
07:50:22 <peter1138> Hmm, I wonder what the latency was on my original 40MB HDD back in 1990. I know the throughput was significantly lower than current broadband speeds...
07:50:42 <andythenorth> Xaroth it's user error :)
07:50:47 <andythenorth> GS is just weird
07:50:53 <andythenorth> far more weird than grf
07:51:07 <TrueBrain> lol, lets agree to disagree andythenorth :P
07:51:11 <TrueBrain> peter1138: you are weird. Someone had to say that
07:53:51 <TrueBrain> someone just finished a 9 hour (!) TURN session :)
07:54:02 <TrueBrain> costs us a wopping 13 MiB of bandwidth
07:54:08 <TrueBrain> OpenTTD is still cute :)
07:54:28 <peter1138> I guess I should update my server :p
07:54:56 <TrueBrain> 2 connections couldn't be established yesterday
07:55:04 <peter1138> Hmm, I miss timestamps on stdout messages. I could make a patch but...
07:56:25 <TrueBrain> so 0.5% of the connections fail. Not a bad value, still higher than I would like :P But I guess we have to await more reports of people who cannot join to investigate :)
07:57:11 <peter1138> But my free time is taken up by moaning about not having gone out cycling.
07:57:23 <TrueBrain> ah, yeah, that is a fulltime job
07:57:48 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9522: Fix: pathfinders always tried to avoid docking tiles (even if nothing was on them) https://git.io/JEieV
08:01:50 <TrueBrain> hmm, why don't we tell on giving of orders that the PF cannot find a route between that order and the one before?
08:02:08 *** Kitrana1 has quit IRC (Quit: Leaving.)
08:05:54 <TrueBrain> I can't be the first to ask, so I assume there is an answer .. either "that is already in there" or "it can't be done because ..." :P
08:06:12 <peter1138> Shared orders and conditional orders could be awkward but probably not insurmountable.
08:06:16 <Xaroth> or "nobody made a patch for that"
08:06:44 <TrueBrain> Xaroth: that is doubtful, giving the age of this problem :D
08:07:02 *** Kitrana has joined #openttd
08:07:05 <peter1138> Is it another bank holiday? The roads were really quiet and my group Teams call is so far just me...
08:07:20 <TrueBrain> every day is a banking holiday, not?
08:07:34 <Xaroth> Well, if peter1138 doesn't have a patch for that, and neither does JGR, the chances of there being patch suddenly gets a lot less likely :D
08:07:52 <TrueBrain> "In former versions the user was informed if a leg was too long when adding a waypoint to the route. So it was obvious that a buoy has to be added." <- euhm .. I don't know all features, but there was a version that did?
08:08:33 <Xaroth> I recall something about being told you need bouys, but not sure if that's OpenTTD or TTD(P)
08:08:42 <LordAro> there used to be a maximum order distance for ships, no? so would've been easy to add
08:08:45 <LordAro> can't remember it in OTTD though
08:09:44 <TrueBrain> it sounds like such an obvious feature :P
08:10:10 <TrueBrain> ha, I will put it on the list of shit I want to do, okay?
08:10:14 <Eddi|zuHause> i'm pretty sure that feature existed
08:10:21 <TrueBrain> I am still waiting for you to fix the languages settings LordAro :P
08:10:48 <peter1138> I know ship order distance got removed.
08:11:01 <LordAro> s'alright, winter is coming, less opportunities to cycle
08:11:07 <peter1138> LordAro, how do I make my legs work better again?
08:11:27 <peter1138> I wonder if riding a sub-10kg bike would help...
08:11:37 <LordAro> it's done wonders for me
08:12:31 <peter1138> The tourer is comfy and convenient but at 16+kg not great for group rides.
08:12:36 <TrueBrain> okay, at least now I understand the reference
08:12:40 <TrueBrain> it was relatively "recent"
08:12:57 <LordAro> peter1138: depends how flat the ride is :p
08:13:18 <LordAro> i'm still struggling with the saddle on my new bike
08:13:24 <LordAro> not very nice to sit on at all
08:15:01 <andythenorth> so how do I string compare in Squirrel?
08:15:13 <peter1138> Bah, turns out I had a 500m elevation gain due to switching from Garmin to Strava mid-ride :(
08:15:14 <andythenorth> do strings have types, like unicode or not?
08:15:47 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened issue #9525: [Bug]: "ship is lost" is only reported after ship leaving depot https://git.io/JESXH
08:15:52 <peter1138> (Garmin ran out of juice, "15 hour battery" lol)
08:15:53 <TrueBrain> there, made it a bug-report :P Is that evil? :D
08:16:17 <peter1138> A ship can't be lost in a depot. It's in the depot!
08:16:21 * andythenorth is so clown shoes
08:16:22 <LordAro> peter1138: impressive
08:16:34 <peter1138> 1030m instead of 1500m :/
08:16:52 <peter1138> First time I've used "Fix Elevation" in Strava.
08:17:19 <andythenorth> I really need to learn how to write makefiles
08:17:24 <LordAro> i always use it, as my device is very unreliable
08:17:30 <andythenorth> makefiles that don't trigger when deps change...make testing very hard
08:17:31 <LordAro> regularly cut 500ft off it
08:17:48 <peter1138> My device is okay but switching devices is not so good.
08:18:01 <peter1138> I wonder if it needs a repair.
08:18:33 <andythenorth> if I have nested foreach()
08:18:49 <andythenorth> I can break in the 2nd nested foreach
08:18:55 <andythenorth> but I need to break the parent
08:18:57 <TrueBrain> closed 1 ticket, opened 2 new ones based on it. I am doing this correct, right?
08:19:15 <andythenorth> do I need to use do..while or something :P
08:19:17 <TrueBrain> andythenorth: in modern languages this is easy, but in Squirrel you would need to keep a boolean
08:19:30 <peter1138> TrueBrain, at least I listed reasons for removal instead of having just the title :D
08:19:38 <andythenorth> shall I just pretend I'm programming javascript in 1998?
08:19:42 <TrueBrain> so before the second foreach, make a boolean set to false. Set it to true before the break, and check after the foreach to break again
08:19:45 <andythenorth> this seems about equivalent
08:19:48 <TrueBrain> or .. use a function for both foreach, and use return
08:20:04 <andythenorth> your first idea was the one I thought of too...it just seems so...wrong :)
08:20:07 <TrueBrain> peter1138: you did, and I was happy with it :D
08:20:20 <andythenorth> ok so Squirrel = Flash Actionscript version 2
08:20:25 <andythenorth> and just don't look too closely
08:20:26 <LordAro> peter1138: what repair? Garmin just tell you to buy a new one, last time i checked
08:20:35 <LordAro> suppose you could try replacing it yourself...
08:20:36 <TrueBrain> andythenorth: these days there is this "for .. else" concept, which allows addressing that differently. But Squirrel doesn't have that :)
08:20:47 * andythenorth ruined by python and even...ugh...jquery
08:21:00 <andythenorth> TrueBrain thanks
08:21:18 <peter1138> LordAro, probably yes :(
08:21:36 * andythenorth detecting grf parameters by building industry fingerprint strings from accept/produce cargos
08:22:10 <peter1138> I didn't even know there was an Edge 130+, with a spec game that makes the estimated battery life more realistic.
08:22:25 <TrueBrain> okay, what else is in 12.0 milestone ..
08:22:43 <TrueBrain> guess andythenorth should fix OpenGFX for us :P
08:22:51 <peter1138> One day I'll continue my work on fixing the UI.
08:22:59 <peter1138> Spacing/padding/scaling/etc/
08:23:04 <TrueBrain> the code will be here when you do :)
08:23:30 <peter1138> I'll probably finish it and then someone will merge a change to Qt or something hideous.
08:23:45 <TrueBrain> the moment we all leave the project, right? :D
08:25:15 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8961: Add: Larger size renderings of the logo/icon https://git.io/JESM0
08:25:29 <peter1138> Also gpsrumors has disappeared, so if I replace the device they'll probably put out a *40-series instead...
08:25:59 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8961: Add: Larger size renderings of the logo/icon https://git.io/JY5D0
08:26:43 <peter1138> Used to have details on life times (release, end-of-sale, end-of life, etc) of various GPS devices, watches, cycle computers, etc.
08:27:11 <peter1138> and predictions of when new devices would come out, using information from FCC filings and such like.
08:30:05 <LordAro> peter1138: the video i just linked claims to be improving the battery life over the original, but they appear to be replacing a 700mAh battery with 600mAh
08:32:19 <TrueBrain> less is more, right? :D
08:32:49 <peter1138> This issue will affect EVs as well... Oh well.
08:35:10 <peter1138> Camden Brewery Marmite Ale... Hmm.
08:36:02 <andythenorth> industry fingerprinting works
08:36:13 <andythenorth> TrueBrain I will fix opengfx, it's only 5 sprites or something?
08:36:16 <andythenorth> for new network UI?
08:36:44 <TrueBrain> no real rush, but if you can, would be much appreciated :D
08:36:53 <TrueBrain> better than solitaire :)
08:38:36 <andythenorth> better than GS though? :P
08:40:25 <andythenorth> I'm not trying for neat diffs here, just 'working'
08:41:04 <andythenorth> this was all instigated by TrueBrain and the idea that GS and Newgrf could be used as 2 APIs from one context :P
08:41:10 <andythenorth> where will it end?
08:41:23 <TrueBrain> wait, am I getting blame here now?
08:41:29 <andythenorth> credit, not blame
08:41:36 <TrueBrain> this is not credit :D
08:42:09 <andythenorth> this is the finest OpenTTD modding for at least 10 years
08:43:19 <TrueBrain> I don't even want to know at this point :P
08:43:47 <andythenorth> I might experiment with simulating industry production in GS
08:43:54 <andythenorth> on a small map, I think I can run it fast enough
08:44:02 <andythenorth> the _actual_ production will be done by grf
08:44:12 <andythenorth> but I can simulate what it would have been in the GS
08:44:20 <andythenorth> on 256 tick or monthly loops
08:44:51 <andythenorth> I think it will be accurate enough for goal / town growth purposes
08:45:11 <andythenorth> I'm now going to template squirrel with python, so I can simulate the behaviour of newgrf industries in GS
08:48:18 <peter1138> Sometimes I hate programming.
08:48:48 <peter1138> Replacing 1 line to 13 because I need to check an extra condition :(
08:49:20 <andythenorth> won't AI do it for us soon?
08:50:58 <TrueBrain> okay, CmdFoundTown misuses _generating_world to work around a bunch of conditions .. fun
08:51:13 <TrueBrain> because of this, other problems happen, ofc
08:51:15 <TrueBrain> as that is how things work :P
08:53:32 <TrueBrain> is there somewhere a variable I can read to know if any station has been build?
08:54:41 <TrueBrain> couldn't find it, because it was in base_station_base.h
08:54:47 <TrueBrain> it is all about that base
08:56:12 <peter1138> andythenorth, so 4 lines of if conditions are now 52 lines. Sadface. I should abstract this out.
08:56:36 <TrueBrain> making it 500 lines of code, you say? :D
08:57:01 <andythenorth> with boilerplate
08:59:18 <andythenorth> and unit tests for your function
08:59:41 <andythenorth> then you could make it a library so it can be reused
09:06:14 <TrueBrain> which requires another repository to be created
09:06:25 <TrueBrain> and it being published on GitHub
09:06:30 <TrueBrain> which requires a maintainer
09:07:02 <andythenorth> and then contribution management
09:07:10 <andythenorth> and security analysis
09:07:19 * andythenorth is enjoying software a lot right now
09:09:56 <peter1138> Maybe I should cycle to the office.
09:10:26 <peter1138> Just to help spread this Covid thing about a bit more.
09:13:12 <andythenorth> is that what we're doing now?
09:20:01 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JESAi
09:21:42 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #9407: [Bug]: CmdFoundTown may cause multiplayer desyncs due to not initialising Town::stations_near https://git.io/JcZJh
09:23:12 <TrueBrain> that was a fun bug :)
09:23:47 <LordAro> can't help but feel like the fix is a bit of a hack also :p
09:24:24 <LordAro> we can't use one global variable, so use a different one instead
09:24:39 <LordAro> why can't generating_world be passed through as a parameter to the function?
09:25:14 <TrueBrain> did you read my PR description? As the point of that if-statement is to prevent an expensive station-lookup. It is an optimization. We can also remove that if-statement, and nothing breaks
09:25:29 <TrueBrain> so why is it a hack to check for the one thing that can influence if that statement does anything?
09:25:43 <TrueBrain> using _generating_world to start with is the hack
09:25:52 <TrueBrain> "as we are sure there are no stations" :P
09:26:24 <TrueBrain> basically I unhack the existing hack by checking the proper information :P
09:26:38 <TrueBrain> I can be convinced otherwise btw, but that is how I understood that if-statement :)
09:27:02 <LordAro> is there no better way to get the number of stations? .items feels a bit internal
09:27:19 <TrueBrain> I dunno :P I just assumed this was the sane way to check :D
09:28:27 <TrueBrain> did expect size() to exist, but this is clearly created "before" :P
09:28:44 <LordAro> there are 2 other usages of .items in story.cpp
09:28:52 <Eddi|zuHause> i think the mistake was to assume there's a sane way :p
09:29:24 <TrueBrain> GetNumItems() sounds sane
09:29:38 <LordAro> i would be tempted to make many of the pools internal variables private/protected
09:30:03 *** virtualrandomnumber has joined #openttd
09:30:16 <TrueBrain> my PR is evidence why
09:30:23 *** virtualrandomnumber has quit IRC ()
09:32:16 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JESAi
09:32:19 <TrueBrain> looks much better LordAro , tnx :)
09:32:49 <LordAro> peter1138: the best i can do is std::synchronized_pool_resource
09:33:02 <LordAro> i am not entirely clear what it is
09:33:04 <LordAro> some sort of allocator
09:35:37 <TrueBrain> the ways we use _generating_world and checks like _game_mode == GM_EDITOR annoy me greatly
09:35:56 <TrueBrain> another global "_skip_town_validation" or something would have been more descriptive :P
09:36:08 * LordAro definitely not baiting TrueBrain
09:36:30 <TrueBrain> yeah, not falling for it (again) :P
09:37:03 <TrueBrain> been there, got the t-shirt, and the completely useless conversation that followed :P
09:37:23 <TrueBrain> but the age of our code shows in these places :) It is fun :)
09:41:13 <TrueBrain> been trying to reverse engineer Legend of Zelda (the SNES one), as I am very curious how they created it back in the days ... I wonder how much weirdness I find there. Dune2 was full of these kind of weirdnesses, that was interesting too :)
09:52:44 <DorpsGek> [OpenTTD/OpenTTD] JGRennison commented on pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JE9f4
09:56:59 <TrueBrain> hmm, that is not a bad idea, to move it inside that function ..
09:57:42 *** murr4y has quit IRC (Quit: WeeChat 3.1)
10:03:01 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JESAi
10:11:05 *** murr4y has quit IRC (Quit: WeeChat 3.2)
10:11:57 *** andythenorth has quit IRC (Quit: andythenorth)
10:23:54 *** andythenorth has joined #openttd
10:26:03 <Timberwolf> TrueBrain: that one always used to break emulators because it used layers/planes in complex ways.
10:26:55 <Timberwolf> So you'd play it and have just the rain and a flat grey land with Link and some trees on it.
10:27:46 <TrueBrain> Yeah .. sprites are hard to render correct
10:27:52 <TrueBrain> Very timing dependent
10:30:19 <peter1138> Much easier on a BBC Micro which didn't have any sprite hardware :D
10:40:28 *** murr4y has quit IRC (Quit: WeeChat 3.2)
10:47:34 <Eddi|zuHause> things that "break emulators" are usually undocumented hardware quirks that just happened to work, so they got used
10:48:39 <Eddi|zuHause> so emulators that just implement the documented parts will break
10:49:26 <Samu> TrueBrain, "In the upcoming nightly, the OP's savegame behaves again as it did before 31db4f8." - that's only true if no ship was at the destination's docking tile.
10:50:03 <TrueBrain> is there a ship at the destination docking tile?
10:50:25 <TrueBrain> you can FF that game for a long time without ever getting "ship is lost"
10:50:30 <TrueBrain> either way, this is a pointless argument
10:50:59 <Samu> anyway, just arguing against the issue being closed
10:51:45 <LordAro> Samu: show concrete examples, not just vague statements
10:52:40 <Samu> the screenshot in the PR
10:53:52 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JE9Xw
10:54:08 <LordAro> Samu: i shall continue to ignore you until you say something actionable
10:54:51 *** andythenorth_ has joined #openttd
10:57:39 *** andythenorth__ has joined #openttd
10:57:39 *** andythenorth_ has quit IRC (Read error: Connection reset by peer)
10:58:08 <LordAro> Samu: so why can't TB reproduce any issues?
10:58:27 <Samu> manually stop it after loading
10:58:34 <Eddi|zuHause> LordAro: i've been doing that for like 5 years now. hasn't happened yet
10:59:07 <LordAro> Samu: see, at no point in the past have you mentioned that
11:00:05 *** andythenorth has quit IRC (Ping timeout: 480 seconds)
11:01:23 <Samu> not arguing against the PR being merged, it fixes something, which is positive, it's just too early to close the issue though
11:02:05 <LordAro> say that in the issue then
11:12:52 <andythenorth__> I miss peter1138 twitch coding
11:15:00 *** andythenorth has joined #openttd
11:20:45 <Samu> it's funny you're quick to complain that I don't test many scenarios, but when you do the same, it doesn't matter
11:21:54 *** andythenorth__ has quit IRC (Ping timeout: 480 seconds)
12:20:00 <peter1138> andythenorth, last time I did that I got "told off" ;-)
12:20:27 <TrueBrain> by who? who should we bitchslap? :P
12:20:40 <peter1138> I won't name names!
12:21:02 <TrueBrain> I always wanted to do a twitch coding stream .. just because it sounds like a lot of fun :P
12:26:14 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #9525: [Bug]: "ship is lost" is only reported after ship leaving depot https://git.io/JESXH
12:28:34 <DorpsGek> [OpenTTD/OpenTTD] PeterN commented on issue #9525: [Bug]: "ship is lost" is only reported after ship leaving depot https://git.io/JESXH
12:31:41 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9526: Fix #9407: desync when founding a town nearby a station https://git.io/JESAi
12:31:44 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed issue #9407: [Bug]: CmdFoundTown may cause multiplayer desyncs due to not initialising Town::stations_near https://git.io/JcZJh
12:39:19 <Samu> on my 15 ais, road vehicle ticks are averaging 62 ms
12:39:33 <Samu> there's about 4000*15 rv's in the world atm
12:40:40 <andythenorth> I don't remember telling peter1138 off :P
12:40:54 <Samu> funny aircraft ticks are higher than ship ticks
12:41:07 <Samu> I guess buoys help a great deal here
12:43:22 <peter1138> andythenorth, no :-)
12:44:42 <peter1138> Samu, do AIs use threads yet?
12:45:10 <TrueBrain> does singleplayer queue commands? :D
12:47:24 *** jottyfan has joined #openttd
12:51:50 <Samu> it's so awesome to see a 4096x4096 becoming populated
12:54:42 <Rubidium> isn't it better to do the minimap export?
12:56:34 <Rubidium> isn't that one much prettier?
12:57:03 <LordAro> the map looks empty unless i actually zoom in :)
12:58:09 <Rubidium> buy a bigger screen ;) Or... the image viewer must be made aware of the content, so it can prioritize the routes over land/water
12:58:33 <TrueBrain> for some reason I imagine "game of life"
13:00:28 <andythenorth> the digital simulation, not the board game
13:01:09 <andythenorth> super patriarchal hetero-normative capitalism
13:01:15 <andythenorth> and also not fun
13:04:39 <Samu> now I understand the power of cache
13:05:08 <Samu> iterating all vehicles is bad, I have a different view on #7918 now
13:06:43 <andythenorth> hmm what was I doing?
13:06:59 <andythenorth> replicating newgrf industry production logic in GS
13:07:15 * andythenorth seriously wondering about an nml -> squirrel transpile
13:07:19 <andythenorth> I'd be so out of my depth :D
13:07:26 <andythenorth> would be very funny
13:09:54 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick commented on pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/JEHLy
13:09:57 <LordAro> do a squirrel to lua transpile first
13:13:09 <TrueBrain> LordAro: you are on a roll, aren't you? :D
13:13:11 <Eddi|zuHause> and then back to squirrel?
13:13:37 <Eddi|zuHause> like those "google translate sings songs" videos?
13:13:57 <LordAro> TrueBrain: it would mean we could write AIs in lua!
13:14:11 <LordAro> while not requiring me to have to rewrite AroAI
13:14:15 <TrueBrain> I wish that actually solved anything, otherwise I would do it :P
13:15:14 <Samu> I need a direction, how to proceed with #7918?
13:15:38 <Eddi|zuHause> TrueBrain: it's like andythenorth switching templating languages every other week
13:18:00 <Samu> question is, do I make a new PR that would superseeds #7918 and implement solution 2 presented in the comments?, or do I overwrite the solution 2 while keeping the same PR? it's going to be more difficult to follow that PR's progress though
13:18:50 <LordAro> Samu: if you're still solving the same problem, overwrite would be preferred
13:19:15 <glx> andythenorth: in french the board game is named "destins", sounds better
13:22:26 <andythenorth> Eddi|zuHause remind me, how many templating languages have I used since 2008?
13:22:45 <Eddi|zuHause> sorry, i lost count :p
13:23:05 <andythenorth> you can't count to 1? :P
13:23:28 <andythenorth> strictly there was C Pre-Processor and variadic macros as well, but I didn't add that
13:23:30 <andythenorth> I only deleted it
13:24:35 <andythenorth> LordAro sqrl -> lua doesn't seem to encapsulate any of grf :P
13:24:55 * andythenorth only wants to transpile the production callback, it can't be that hard?
13:25:41 <andythenorth> it's just a bunch of if statements
13:25:44 <Eddi|zuHause> that's simple, you just have to add an output module to nmlc
13:26:05 <andythenorth> we've got the AST...
13:27:31 <andythenorth> I suspect I'll have to manually recreate it
13:27:50 <andythenorth> simply taking the logic won't work, there'll have to be abstractions for things that are in permanent storage
13:28:00 <andythenorth> and maybe for vars that GS can't read
13:28:08 <andythenorth> but I think it's viable
13:29:09 <Eddi|zuHause> you have to map permanent storage to a squirrel data structure, and add no-ops for the unreadable vars
13:30:14 <Eddi|zuHause> you can add simplifications for that assuming you only want to output partial programs, not complete runnable ones
13:30:15 <andythenorth> storage in this context is populated by operations on amounts of cargo received
13:30:25 <andythenorth> which GS can access, but not in the same time frame AFAICT
13:30:42 <andythenorth> so there needs to be some adjustment to production calculation for timespan
13:30:52 <andythenorth> eh, it's only a simulation of the industry :)
13:32:42 *** Speeder__ has joined #openttd
13:40:00 *** Speeder_ has quit IRC (Ping timeout: 480 seconds)
14:23:40 <Samu> autoreplace and vehicle group management is ... really hard to follow in code
14:23:53 <Samu> i have to think there's 2 vehicles temporarily
14:24:45 <Samu> the old vehicle is still in existance in the group where the new vehicle is getting put into. this confuses statistics copying
14:40:11 *** Feuersalamander has joined #openttd
14:42:09 *** grossing has quit IRC (Ping timeout: 480 seconds)
14:43:41 *** azubieta60 has quit IRC (Quit: Ping timeout (120 seconds))
14:44:08 *** azubieta60 has joined #openttd
14:45:20 *** avdg has quit IRC (Ping timeout: 480 seconds)
14:45:45 *** duckfullstop has joined #openttd
14:45:45 <Samu> I don't have a good way to test this fully :(
14:46:20 *** luaduck has quit IRC (Read error: Connection reset by peer)
14:46:51 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/Jvek6
15:06:13 *** gelignite has joined #openttd
15:17:58 <Samu> just updated main topic with more text!
15:19:39 <DorpsGek> [OpenTTD/OpenTTD] BadBrett opened issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
15:21:28 <FLHerne> wait, is that THAT Bad_Brett?
15:21:43 <LordAro> can't be that many of them
15:22:08 <FLHerne> Wow, he's active again on the forum, yes
15:22:52 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
15:24:35 <LordAro> hmm, game seed does not appear in the gamelog
15:30:25 *** jottyfan has quit IRC (Quit: jottyfan)
15:30:40 <TrueBrain> You need more than the seed I think .. also all terrain settings
15:31:15 <LordAro> still, can probably place it on the border manually
15:31:44 <TrueBrain> (See what I did there?)
15:35:16 <LordAro> the objects in the GRF provided cannot be manually placed
15:36:53 <LordAro> oh hang on, it's a desert object
15:37:01 <LordAro> of course it is, it's Bad_Brett
15:40:26 <LordAro> hmm, object size is 0x0
15:40:31 <LordAro> this feels familiar..
15:43:12 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
15:44:18 <FLHerne> I feel like I did that already
15:44:40 <LordAro> thought it was familiar
15:44:55 <LordAro> though there is a size: [2,2]; in the nml file provided
15:46:38 *** Etua has quit IRC (Quit: Etua)
15:46:46 <FLHerne> Oh, mine was 8166, a roadtype with no label
15:47:10 <FLHerne> I think I was half-remembering 8276 also, but that's dp's
15:52:10 *** sla_ro|master has quit IRC ()
16:07:44 *** Progman has joined #openttd
16:18:05 *** Wormnest has joined #openttd
16:18:55 <DorpsGek> [OpenTTD/OpenTTD] BadBrett commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
16:21:09 <TrueBrain> (tile=65569) <- lol .. LordAro , I guess you are on a 256x256 map? :D
16:21:32 <TrueBrain> ah, yes, that is in there too
16:23:09 <LordAro> (because the object is apparently being (partly?) recognised as 0x0, i can't place it manually either
16:23:31 <TrueBrain> I guess it works for 1.11.2 for example because asserts are disabled :P
16:23:46 <LordAro> that'd certainly help
16:24:01 <TrueBrain> seems CmdBuildObject doesn't do any bound checking :P
16:24:09 <TrueBrain> so we can play: this has been broken for how long? :D
16:24:23 <LordAro> assuming that's where the bounds checking should be
16:24:31 <LordAro> probably as long as NewObjects
16:25:00 <TrueBrain> well, TileArea doesn't do bound checking either, it seems
16:26:25 <TrueBrain> int size_x = GB(spec->size, HasBit(view, 0) ? 4 : 0, 4);
16:26:28 <TrueBrain> do I even want to know?
16:27:17 <LordAro> objects store their size in a single byte
16:27:37 <TrueBrain> yeah, but the flipping of w/h .. it is amusing :)
16:30:02 <LordAro> ¯\_(ツ)_/¯ i didn't make the spec :p
16:30:30 <TrueBrain> I am not pointing fingers :) Just NewGRF doesn't stop amazing me :)
16:31:20 <TrueBrain> ....... the moment I hate unicode ...
16:31:38 <TrueBrain> give me back the time it was just being rendered as a box with a cross in it plz
16:32:26 <TrueBrain> okay, that is just plain awesome
16:34:54 *** J0anJosep has joined #openttd
16:38:56 <J0anJosep> I get this error message in the commit checker for PR#8480: No common parent found for this merge commit (max-depth of 256 reached)
16:39:22 <J0anJosep> Any idea of how I can solve it?
16:40:45 <glx> usually means you're on the wrong branch when opening the PR
16:43:20 <glx> but it doesn't seem to be the case here
16:45:06 <glx> maybe a bug in commit checker (75 commits is not usual ;) )
16:47:54 *** frosch123 has joined #openttd
16:48:06 <J0anJosep> I think it passed the tests before with more than 80 commits.
16:50:57 <FLHerne> OpenTTD proper didn't do the master -> main change yet, did it?
16:52:22 <FLHerne> TrueBrain: My favourite grf thing is the way all offsets are relative to the *invisible* corner of the bounding box
16:52:56 <FLHerne> Like, if it was the front 'middle' corner' you could count pixels from it in the offset tool
16:53:24 <FLHerne> but no, all the offsets are relative to the one corner that's not displayed at all
16:54:32 <TrueBrain> we need to write a book about it :P
16:54:36 <frosch123> FLHerne: you did not get to child sprites then :p
16:54:37 <TrueBrain> oddities in OpenTTD / NewGRF
16:56:15 <andythenorth> my NEW favourite thing is 15 bit callback results
16:56:48 <TrueBrain> J0anJosep: honestly, I wouldn't worry about that CI part failing; if you are done with your PR and are happy with it, we need to talk how we are going to merge that all anyway :)
16:56:53 <TrueBrain> and I doubt that would be in a single PR ;)
16:58:56 <andythenorth> maybe there is a frog
16:59:05 <andythenorth> the subtext could be 'quak'
16:59:11 <TrueBrain> hmm, elements updated, and tab-completion is now annoying :P Yay for progress!
16:59:37 <frosch123> andythenorth: interesting site
16:59:43 <TrueBrain> glx: guess that "=" should be "-gt" :D
17:00:02 <TrueBrain> so yeah, 64 commits is currently the limit :P
17:00:58 <J0anJosep> TrueBrain: Ok. I won't worry about it then.
17:03:18 <glx> I don't plan to buy a new computer so
17:03:37 <TrueBrain> I wonder if that gives any new crashes
17:05:08 <TrueBrain> this will be the rest of the evening now, not?
17:06:07 <frosch123> TrueBrain: windows 11 still does not run on leap days
17:07:08 <J0anJosep> Is it possible to create a preview for a PR I have in my own repository?
17:07:33 <TrueBrain> nope; you cannot read the secrets required for that
17:07:34 <TrueBrain> you would need to setup your own AWS account if you like to do that
17:07:37 <TrueBrain> I wouldn't advise it :)
17:07:43 * Rubidium wonders whether buying a computer in a time of huge chip shortages is the best time... seems a bit... uhm... weird
17:09:45 <J0anJosep> TrueBrain: Thanks! And can I prevent the CI being executed in my repository? It runs every time I push a PR and I get emails each time.
17:09:51 <DorpsGek> [OpenTTD/actions] TrueBrain opened pull request #34: Fix: [checkout-pull-request] PR commit limit kicked in one round too early https://git.io/JEQeH
17:10:04 <TrueBrain> J0anJosep: under actions, click the dots, click disable
17:10:22 <TrueBrain> well, click the workflow first
17:10:32 <TrueBrain> than the dots right of the "filter workflow runs" text thingy
17:10:52 <J0anJosep> TrueBrain: Thank you!
17:10:54 <TrueBrain> it won't run that workflow again till you enable it
17:14:37 * andythenorth saying random words
17:15:20 <glx> TrueBrain: yup ran the script by hand locally, that's definitely the issue
17:18:18 <TrueBrain> good! Review the PR, and lets gooooo :)
17:18:31 <TrueBrain> I hate releasing a new set of actions, requires me to manually do stuff on the CLI :P
17:19:01 <DorpsGek> [OpenTTD/actions] glx22 approved pull request #34: Fix: [checkout-pull-request] PR commit limit kicked in one round too early https://git.io/JEQfY
17:19:28 <DorpsGek> [OpenTTD/actions] TrueBrain merged pull request #34: Fix: [checkout-pull-request] PR commit limit kicked in one round too early https://git.io/JEQeH
17:20:32 *** duckfullstop is now known as luaduck
17:22:28 <TrueBrain> lets see if it actually works ... retriggered the job
17:23:12 <glx> anyway even with windows 11 I'll still need a 32bit XP VM to use my scanner :)
17:24:27 *** HerzogDeXtEr has joined #openttd
17:24:29 <glx> commit checker is happy now
17:25:43 <glx> I think the biggest PR tested with it was CMake
17:25:52 <TrueBrain> that was before I moved it :P
17:26:23 <glx> it was included in workflow (I wrote the script)
17:26:54 <glx> remember it was full of bashism on first try
17:27:10 *** Progman has quit IRC (Remote host closed the connection)
17:27:18 <TrueBrain> no no, I recently converted the TypeScript version to bash again, remember? :D
17:27:24 <TrueBrain> pretty sure this happened while doing that
17:27:50 <TrueBrain> and the reviewer didn't catch it
17:27:54 *** Progman has joined #openttd
17:29:49 <glx> ah it wasn't me for this one, (but I would probably have missed it)
17:30:03 <TrueBrain> did you really look up who to blame? Lol :P
17:33:36 <frosch123> stringids are funny, STR_CONFIG_SETTING_PATHFINDER_YAPF_RECOMMENDED <- adding the "recommended" into the stringid makes it harder to change the recommendation? :p
17:33:53 <TrueBrain> yeah ... there is a lot of ... "fun" things in there
17:34:14 <TrueBrain> I most hate the place we put "TRAIN" / "ROAD" / etc .. that is just unneeded more complex
17:35:17 <frosch123> so, was there no unused string in the end?
17:36:06 <TrueBrain> I expected a lot more :P
17:36:08 <frosch123> ah, different PR :)
17:37:20 <TrueBrain> yeah .. required a bit of archeology to figure out what happened for some :P
17:43:07 <frosch123> oh dear, you had to call cpp as subprocess :) don't give andy any ideas
17:43:37 <TrueBrain> so why mention it! :P
17:49:06 <Rubidium> frosch123: if you want it changed, add it to #9101 ;)
17:51:26 <frosch123> wow, you expect me to remember something i said 16 minutes ago, and put things into context? :p
17:51:51 <frosch123> i wondered 2 minutes how 9101 changes things for cpp :p
17:52:44 <TrueBrain> he is a bit slow :P
17:53:35 <frosch123> but i'll pass it on: TrueBrain, if you want to change the _TRAIN strings etc, put them into #9101 :p
17:54:13 <TrueBrain> I was just writing that
17:55:55 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on discussion #9101: Which strings in english.txt should be renamed or moved before 1.12? https://git.io/JOQOA
17:57:26 <glx> yes makes sense to put the type at the end
17:57:30 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on discussion #9101: Which strings in english.txt should be renamed or moved before 1.12? https://git.io/JOQOA
17:57:53 <glx> most string id are based on the actual string text
17:58:08 <TrueBrain> there frosch123 , even added the RECOMMENDED
17:58:26 <glx> but it's not the best for automatic stuff
17:59:12 <frosch123> TrueBrain: i'll spread myths about truepf
17:59:15 <TrueBrain> what do you mean glx?
17:59:25 <TrueBrain> frosch123: haha, that would have been better than yayapf :)
18:00:26 <glx> STR_NEWS_FIRST_TRAIN_ARRIVAL :{BIG_FONT}{BLACK}Citizens celebrate . . .{}First train arrives at {STATION}!
18:00:26 <glx> STR_NEWS_FIRST_BUS_ARRIVAL :{BIG_FONT}{BLACK}Citizens celebrate . . .{}First bus arrives at {STATION}!
18:01:32 <TrueBrain> later strings show it is not really related to the actual string
18:01:41 <TrueBrain> but somehow this sense of writing down stringids just kept going
18:01:46 <TrueBrain> "_TOOLTIP" MUST COME LAST
18:02:04 <glx> yeah new strings are more smart for the id
18:02:21 <glx> but as always the legacy :)
18:15:12 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #9518: Codechange: validate what strings are defined but not used https://git.io/JEQYl
18:22:18 <frosch123> looks like the release of ottd's new networking has caused china to limit online games to 3 hours/week for children
18:23:14 <andythenorth> my kids are horrified by that news
18:23:21 <andythenorth> they see it as infringement of basic rights
18:23:30 <andythenorth> someone now has to say "and this is why China will win"
18:25:46 <Eddi|zuHause> remember back in the old days, parents tried to limit TV to 1h per day...?
18:28:07 <DorpsGek> [OpenTTD/actions] J0anJosep commented on pull request #34: Fix: [checkout-pull-request] PR commit limit kicked in one round too early https://git.io/JEQ3q
18:31:56 <andythenorth> my parents just refused to own a TV
18:32:02 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:33:08 <Eddi|zuHause> mine tried that too, for like half a year
18:33:30 <frosch123> i don't have a "tv"
18:33:39 <frosch123> sounds like last century
18:34:25 <Eddi|zuHause> there's an xkcd for that
18:35:04 <andythenorth> is there an xkcd for "there's an xkcd for that"?
18:36:40 <frosch123> rule 33? if it exists, there is an xkcd about it?
18:36:55 *** sla_ro|master has joined #openttd
18:46:57 <Eddi|zuHause> i have a satellite card in my pc, but i used it maybe like 3 times in the last 2 years
18:55:53 <peter1138> I had a DVB-T USB stick somewhere, but we moved on to DVB-T2...
18:57:29 <frosch123> dvb-t2 feels like fat32
18:57:36 <frosch123> already obsolte on release
18:59:44 <andythenorth> I have youtube...
19:00:21 <DorpsGek> [OpenTTD/OpenTTD] PeterN updated pull request #9289: Change: Shortcut varaction chains for callbacks. https://git.io/JsHuK
19:09:57 <Xaroth> Rubidium: I have an example admin port thing for #9456 , will comment once I've cleaned this up a bit... also found a bug with the admin port while making this....
19:10:25 *** roadt has quit IRC (Ping timeout: 480 seconds)
19:15:56 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #9528: Fix a2051bad: SendCmdNames doesn't fill up a packet as much as possible https://git.io/JEQCx
19:16:19 <TrueBrain> I love 1 character fixes :)
19:16:52 <TrueBrain> I welcome suggestions for better commit messages
19:19:46 <frosch123> remove the negation: SendCmdNames only sent one name per packet
19:20:01 <DorpsGek> [OpenTTD/OpenTTD] frosch123 approved pull request #9528: Fix a2051bad: SendCmdNames doesn't fill up a packet as much as possible https://git.io/JEQWg
19:20:06 <glx> I'm trying to understand the "more data" bool in comment, and github makes me think there's an alignment issue
19:20:18 <TrueBrain> yeah, GitHub doesn't like tabs anymore
19:20:20 <TrueBrain> it annoys me greatly
19:20:39 <frosch123> glx: the send_bool after the loop
19:20:39 <TrueBrain> tnx frosch123, that is a lot better indeed :D
19:21:22 <TrueBrain> it strongly assumes TCP, this packet :)
19:22:06 <frosch123> to append a string you need enough room for send_bool+send_uint16+send_string+0-term+send_bool after loop
19:23:03 <glx> my brain is just confused by "more data" and "no more data" bools
19:23:27 <TrueBrain> it is more: "last-entry-in-this-packet"
19:23:38 <frosch123> no idea why they are needed :) the packet-length would also have worked
19:23:40 <glx> but false is nothing follows in this packet
19:23:46 <TrueBrain> well: not-last-entry-in-this-packet, ofc
19:24:12 <TrueBrain> it doesn't even tell you if you can expect more entries in another packet
19:24:16 <TrueBrain> which is just the best :)
19:24:17 <glx> while true means a command (and a bool) follows
19:25:00 <frosch123> [21:24] <TrueBrain> it doesn't even tell you if you can expect more entries in another packet <- actually, that makes a lot more sense. maybe someone intended that and messed up the true/false
19:25:23 <TrueBrain> I would have expected an enum ..
19:25:33 <TrueBrain> ENTRY_FOLLOWS, PACKET_FULL, END_OF_LIST
19:25:54 <TrueBrain> but, I guess, this works too ..
19:26:15 <glx> it works because it's not UDP :)
19:27:10 <DorpsGek> [OpenTTD/OpenTTD] Xaroth commented on pull request #9456: Fix #9388: thread unsafe use of NetworkAdminConsole/IConsolePrint https://git.io/JEQlr
19:27:13 <frosch123> SendProtocol does the same weird true/false thing
19:27:30 <TrueBrain> once one person starts with it ... copy/paste is not far behind :)
19:28:21 <TrueBrain> frosch123: that one is even more weird .. why not just start with a length word or something :)
19:28:46 <TrueBrain> or use the ID for it, with 0xffff or what-ever :)
19:31:06 *** jottyfan has joined #openttd
19:33:14 <glx> seems the number of commands is not even sent
19:33:57 <TrueBrain> that isn't bad an sich .. there is this debate about how to send stuff over network .. length-prefixed, or a "there-is-no-more" marker at the end
19:34:03 <TrueBrain> from strings to complete packets
19:34:42 <TrueBrain> and mostly people are strong believers of one camp over the other, as that goes :P
19:34:59 <glx> but in the particular case of these packets, it's none of both
19:35:24 <TrueBrain> that is what is odd about it :)
19:36:08 <glx> receiver will just get an unknown number of packets and luckily it's not UDP
19:36:29 <TrueBrain> unknown number of packets with an unknown number of entries for the same list :)
19:36:32 <TrueBrain> at least it is a dict
19:36:35 <TrueBrain> so it is fiinnnneeeeee
19:37:09 <TrueBrain> the Game Coordinator sends the server-list also in multiple packets, without telling how many to expect exactly
19:37:40 <frosch123> repo is on dead devzone :p
19:37:44 <TrueBrain> each packet starts with how many servers are in there
19:37:45 <TrueBrain> and the final packet is one with servers=0
19:38:22 <TrueBrain> frosch123: sounds about right
19:38:43 <glx> a adminport API implementation
19:38:46 <frosch123> that's the first/reference admin port client
19:39:00 <frosch123> written by the same dih who implemented the stuff in ottd
19:39:12 <TrueBrain> ah, so we have someone to blame :P
19:39:21 <TrueBrain> I keep waiting for the moment someone says: owh, it is already done this way in the game protocol
19:39:25 <TrueBrain> as that most likely means my name is on it :P
19:39:39 <frosch123> so, looks like joan will ignore any data after the "false"
19:39:49 <TrueBrain> sweet, side channel!
19:39:55 <frosch123> and range-except if the package ends without a false
19:39:58 <glx> it will be retriggered by next packet
19:40:38 <glx> but yeah false is end-of-packet
19:40:41 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9528: Fix a2051bad: SendCmdNames only sent one name per packet https://git.io/JEQCx
19:41:05 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/JEQ4q
19:41:33 <TrueBrain> it is just pure evil: this packet is finished, and I will not tell you if you can expect any more of the same type
19:41:50 <TrueBrain> I just hear someone gniffiling every time this packet is used
19:42:02 <TrueBrain> gniffiling .. is that valid English?
19:42:07 <TrueBrain> dunno .. you get what I mean :P
19:42:13 <glx> forcing handling packets as events
19:42:22 <TrueBrain> more than a chuckle
19:43:01 <frosch123> hmm, receiveServerCmdNames is "synchronised", but then it stores the data in threadlocal storage
19:43:24 <TrueBrain> awh, the rabbit hole wasn't nice and warm?
19:43:36 <TrueBrain> well, honestly, it is Java .. it was very very hot ... hellish, I dare say :P
19:44:01 <glx> (at least it's consistent)
19:44:24 <TrueBrain> I still remember one of my teachers at the uni, after I made a remark of having a strong distaste for Java: "you will need to learn to like Java, as you cannot find an IT job later if you cannot write Java"
19:45:00 <TrueBrain> such a cute remark :)
19:45:06 <TrueBrain> he wasn't even that old or anything ..
19:45:29 <Rubidium> that isn't as bad as... "write comments, otherwise the compiler does not know what it is supposed to do"
19:45:49 <TrueBrain> who-ever said that should no longer be teaching what-ever it was he was teaching
19:46:07 <TrueBrain> there is a difference between "onbekwaam" and just being naïve :)
19:46:33 <Xaroth> I had to endure classes for Novell's stack, teacher kept saying how it was the next big thing and you'd be able to land a job at any big firm
19:46:52 <TrueBrain> you make it sound like you are 50 :P
19:46:58 <glx> isn't novell dead with nt4 ?
19:47:20 <Xaroth> It was already dead at that time, yes
19:47:21 <Rubidium> nah, not if you used Windows 9x
19:47:33 <_aD> Xaroth: Ironically, there are probably a few big firms which are still offering the big bucks for Novell greybeards...
19:47:49 <TrueBrain> for all the wrong reasons ....... :P
19:47:50 * Rubidium actually set up a Novell network back in the day
19:48:08 <Rubidium> but alas... token ring and the likes
19:48:17 <Rubidium> anyone know what happened to 3com?
19:48:27 <_aD> I just managed to appear as Ethernet was de-facto, thankfully...
19:49:00 <frosch123> _aD: i had job interviews at software ag (adabas, natural) and some insurance company (PL/I), just for fun :p
19:49:08 <glx> I have a 3com card somewhere, never managed to make it work
19:49:29 <frosch123> the former called "natural" the "better cobol" :)
19:49:32 <glx> and it was a plug and play one
19:50:04 <glx> (well win9x plug and play ;) )
19:50:08 <_aD> frosch123: Organic Cobol (TM)
19:54:47 <Rubidium> Xaroth: thanks for the code, I hope I have time to fiddle with it next month
19:54:57 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick commented on pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/JEQB1
19:55:03 <frosch123> looks like ipx has 32+48 address bits, so way more than ipv4
19:56:03 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
19:58:05 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/Jvek6
19:58:12 <Xaroth> Rubidium: you know where to find me if you get stuck :)
19:59:01 <frosch123> be careful, maybe rb is in a timezone, where it's already 2021-09-01
20:04:01 <_aD> FIRS Steeltown: Go big or go home.
20:04:12 <_aD> I think once I hit MagLevs I'm going to go scorched earth.
20:09:33 <DorpsGek> [OpenTTD/OpenTTD] BadBrett commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
20:12:46 <FLHerne> Rubidium: ...you know that in NML's parser, the comments are required for the compiler to know what it does? :p
20:13:06 <FLHerne> well, I suppose docstrings aren't syntactically comments
20:14:04 <frosch123> that's ply's idea, not nml's, is it?
20:14:57 <glx> nice use of python features
20:15:02 <Samu> michi_cc, I've been thinking... I suspect the move I just did is gonna be worse. If 'copy group membership' fails, the viewports have already been switched! And that's undoable
20:17:06 *** virtualrandomnumber has joined #openttd
20:17:26 *** virtualrandomnumber has quit IRC ()
20:20:08 <Samu> is there a quick way to force a command to fail?
20:21:04 <glx> make it return CMD_ERROR I guess
20:21:43 *** J0anJosep has left #openttd
20:27:03 <Samu> thx, seems that I'm mistaken
20:27:37 <Samu> apparently these viewports changes didn't happen, or I didn't notice anything different
20:30:48 <Samu> when i tested this, that block wasn't ever skipped, now it is for some reason
20:32:18 <Samu> nevermind, i can't use CMD_ERROR anywhere I like
20:33:43 <Samu> TEST and EXEC, I keep forgetting about them
20:34:16 *** gelignite has quit IRC (Quit: Stay safe!)
20:38:10 <Samu> yeah, if it passes in TEST mode, it won't ever fail in EXEC mode. I tried making it fail in TEST mode, but that resulted in the viewport changes to never happen
20:38:35 <Samu> very safe code measures you guys have
20:39:06 * andythenorth did a class in pascal
20:39:14 <andythenorth> but I skipped the class, just did the assignment
20:39:40 <andythenorth> got some comment about 'achieves the result, but is top-down design' or something similar
20:42:02 <andythenorth> I had no idea what that meant at the time, due to skipping the lectures :P
20:44:48 <Samu> just re tested against the assert. Everything's fine, no assert happened. Costs between gui and group still match!
20:45:12 <Samu> sorry for the false alarm michi_cc
20:46:14 <DorpsGek> [OpenTTD/OpenTTD] LordAro opened pull request #9529: Fix #9527: Crash when trying to place multitile objects at map edge https://git.io/JEQ2c
20:46:20 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
20:48:28 <frosch123> what was not correct about the non-zero size check?
20:48:41 <LordAro> it only masked, didn't bit shift
20:48:55 <LordAro> (and then checked for 0)
20:48:57 <glx> did you check NFO output compared to spec ?
20:49:25 <LordAro> glx: i have not touched grf, because i do not know what i am doing
20:50:10 <frosch123> i don't mind changing it to GB, but it's just a codechange, no fix :)
20:50:21 <LordAro> no, it's definitely a fix
20:51:02 <DorpsGek> [OpenTTD/OpenTTD] James103 commented on pull request #9529: Fix #9527: Crash when trying to place multitile objects at map edge https://git.io/JEQ2x
20:51:05 <LordAro> `0x20 & 0xF0` gets read as 32, not 2
20:51:06 <michi_cc> If you compare to 0, shifting isn't doing aynthing.
20:52:36 <LordAro> i did get the directions the wrong way round in my comment however
20:54:18 <frosch123> it's weird that OrthogonalTileArea has a ClampToMap(), but no check for inside-map
20:54:22 <glx> let me compile the grf to check nfo
20:55:02 <LordAro> i couldn't find any other examples of TileArea being clamped either
20:55:11 <LordAro> but i might not have been looking hard enough
20:56:16 <frosch123> LordAro: 0 << 4 is still 0
20:56:56 <LordAro> ...i am not trying to bitshift in that direction
20:57:58 <glx> but should not change == 0 result
20:58:30 <frosch123> 0 stays 0 no matter in what direction you shift
20:58:50 <frosch123> you keep giving the example with comparing to 2, but you are not comparing to 2
20:58:56 <LordAro> 0x22 - 2x2, in byte form. i want to check that neither dimension is 0. Previously, it only did `s & 0xF0`, which only zeros out the lower 4 bits. This means you were left checking 0x20 == 0, which is not helping
20:59:23 <frosch123> well, read the second part of the line
20:59:46 <LordAro> i know, i added the thing
21:00:00 <frosch123> maybe get some sleep :) i don't think irc is the right place to explain hex
21:00:07 <LordAro> i'm having difficulty spelling this out further
21:00:26 <frosch123> i'll quote you on that tomorrow, when you are awake :)
21:00:36 <michi_cc> 0x20 != 0, just like 0x02 != 0, while 0x00 == 0 just like 0x0 == 0.
21:01:38 <glx> you inverted 0xF0 and 0x0F tests during the GB move, was that the wrong part ?
21:02:51 <glx> oh 0x02 could pass while it should not
21:05:41 <DorpsGek> frosch123: Error: unsupported operand type(s) for &: 'float' and 'float'
21:06:32 <frosch123> so, how does 0x02 pass?
21:07:02 <glx> hmm but then 0x20 would be an issue ;)
21:07:18 <DorpsGek> frosch123: Error: unsupported operand type(s) for &: 'float' and 'float'
21:07:40 <glx> the new code does the same in the end, it's a ||
21:08:29 <glx> the only difference is the meaning of the thing compared to 0
21:08:52 <frosch123> it's fine for a codechange :)
21:08:56 <glx> but 0x20 == 0 and 0x2 == 0 are both false
21:08:57 <TrueBrain> I think nobody is doubting that 2 != 32, but why 32 == 0 is different from 2 == 0
21:10:00 <LordAro> regardless of whatever the hell is going on, before i changed the code that if block was not getting triggered for this test grf
21:10:24 <LordAro> so it was still using the "broken" 0x02 / 0x20 size
21:10:32 <LordAro> after, it was using the default 0x11
21:11:14 <glx> weird, I see no reason for that in old or new code
21:14:46 <frosch123> i'll stop now, it's too silly to discuss
21:16:19 <TrueBrain> Msys bug? Would be cool :D
21:18:38 <glx> 0C prop is 22 in NFO, nmlc output is correct
21:22:49 <Rubidium> (or basically any -Ox except -O0)
21:24:36 <glx> crash_test.grf doesn't set prop 0C at all, so it's expected to not enter this block
21:24:44 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:26:41 <LordAro> have i confused the transmitter object for the rocks object?
21:26:55 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on issue #9527: [Crash]: NewGRF with multi-tile object causes OpenTTD 12.0-beta2 to crash on map generation https://git.io/JEH61
21:33:47 <glx> anyway regarding #8276, nmlc should error if object size is not set
21:34:13 <LordAro> there's already an issue about it :p
21:35:03 <glx> well I have a "broken" nml file, I should trace it
21:35:50 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
21:40:42 <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9529: Fix #9527: Crash when trying to place multitile objects at map edge https://git.io/JEQ2c
22:04:01 *** andythenorth has quit IRC (Quit: andythenorth)
22:09:00 *** Samu has quit IRC (Ping timeout: 480 seconds)
22:33:13 *** sla_ro|master has quit IRC ()
22:45:00 *** Progman has quit IRC (Remote host closed the connection)
23:40:44 *** LouWestin has joined #openttd
23:57:09 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8719: Fix #8276: Crash when a NewGRF object's size was not set https://git.io/JEQ5N
23:58:16 <glx> reading asm is fun, but I still prefer high level languages
continue to next day ⏵
