Log for #openttd on 16th September 2021:
Times are UTC Toggle Colours
00:03:40  *** WormnestAndroid has quit IRC
00:03:47  *** WormnestAndroid has joined #openttd
00:07:42  *** Gustavo6046 has quit IRC
00:07:44  *** Gustavo6046 has joined #openttd
00:59:55  *** Fuco has quit IRC
01:44:34  *** _aD has quit IRC
02:15:44  *** debdog has joined #openttd
02:19:05  *** D-HUND has quit IRC
02:41:46  *** Wormnest has quit IRC
02:56:22  *** glx has quit IRC
04:43:27  *** Flygon has joined #openttd
04:57:01  *** WormnestAndroid has quit IRC
04:57:06  *** WormnestAndroid has joined #openttd
05:01:11  *** crem has quit IRC
06:30:43  *** crem has joined #openttd
06:37:37  *** sla_ro|master has joined #openttd
06:38:00  <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on discussion #8397: Daylength
06:39:14  *** Gustavo6046 has quit IRC
06:40:07  *** Gustavo6046 has joined #openttd
06:44:09  *** Gustavo6046_ has joined #openttd
06:48:15  *** Gustavo6046 has quit IRC
06:48:39  *** Gustavo6046_ is now known as Gustavo6046
06:54:12  *** Gustavo6046_ has joined #openttd
06:57:35  *** Gustavo6046 has quit IRC
06:57:49  *** Gustavo6046_ is now known as Gustavo6046
07:02:14  <DorpsGek> [OpenTTD/team] LordAro closed issue #249: [hu_HU] Translator access request
07:03:08  <DorpsGek> [OpenTTD/team] LordAro commented on issue #249: [hu_HU] Translator access request
07:04:43  <DorpsGek> [OpenTTD/team] LordAro commented on issue #251: [ru_RU] Translator access request
07:04:50  <DorpsGek> [OpenTTD/team] LordAro commented on issue #253: [ga_IE] Translator access request
07:27:19  *** HerzogDeXtEr has quit IRC
08:07:56  *** katp32 has joined #openttd
08:08:52  <katp32> heyo, by default the interface is waayy too small to be usable for me, but if I set it to double size, then it also makes the cursor comically large. is there a way to change the interface size without affecting the size of the cursor..
08:15:01  <LordAro> none that i'm aware of, i'm afraid
08:15:15  <LordAro> what do you class as "comically large" ? usually it matches the interface quite well
08:16:31  <katp32> at normal scaling it's fine, but at double scaling the mouse cursor is ~twice as large as the interface buttons, and at 4x scaling the mouse is so large that it becomes difficult to even click the buttons due to the fact that it obscures the entire thing
08:17:46  <LordAro> seems alright to me
08:17:50  <_dp_> katp32, you can try to increase font size as well to increase buttons
08:17:53  <LordAro> are you increasing the font size at the same time as the interface?
08:18:14  <katp32> no, not changing the font size at the same time
08:18:20  <LordAro> try that :)
08:18:25  <LordAro> otherwise post a screenshot of what you're seeing
08:18:52  <_dp_> though ingame cursor is about twice as large as the system cursor for me as well
08:18:56  <_dp_> but I'm used to it by now
08:19:24  <katp32> yea, I wish the cursor just stayed the same as the system cursor...
08:19:57  <katp32> cursor capture also doesn't seem to work on Wayland but that's not a big deal
08:20:58  <LordAro> the cursor has several different variants, which makes just using the system cursor non-ideal
08:21:01  <LordAro> cursor capture?
08:21:31  <katp32> making the cursor not actually move when dragging the viewport (and when turned on in settings)
08:21:54  <LordAro> curious
08:22:49  <LordAro> i'd guess an SDL bug
08:22:55  <katp32> possibly
08:23:27  *** Samu has joined #openttd
08:25:31  *** andythenorth has joined #openttd
08:25:45  <andythenorth> I'm wrong and you backlog defenders are right :)
08:25:47  <andythenorth> (at the end)
08:26:15  * andythenorth tends to pay attention to Poppendiek writing
08:28:34  <LordAro> katp32: regardless, would appreciate a proper bug report so that it can be investigated
08:28:38  <LordAro> as this is the first i've heard of it
08:33:04  *** felix has quit IRC
08:34:46  *** katp32 has quit IRC
08:44:42  *** felix has joined #openttd
08:53:45  *** andythenorth has quit IRC
09:01:55  <Samu> hmm my AI is leaving stopped ships in depots for years, I failed somewhere
09:10:16  *** Etua has joined #openttd
10:09:15  <FLHerne> andythenorth[m]: hm, someone agrees
10:09:44  <FLHerne> maybe this tagging thing is the best of both worlds?
10:11:02  <FLHerne> we get a list of bugs people have cared to report, and also a list of ones tagged (priority: high) or whatever that are the ones that are actually important short-term
10:56:06  <Samu> I think I spotted the problem!
10:56:20  <Samu> vehicle_ready_to_start was not being set properly
11:01:53  <Samu> confirmed! problem really solved! yay
11:13:22  *** Fuco has joined #openttd
11:14:33  <peter1138> Oh no, I need to come up with project timescales :(
11:15:38  *** WormnestAndroid has quit IRC
11:16:59  *** WormnestAndroid has joined #openttd
11:57:14  *** Fuco has quit IRC
12:13:14  *** tokai|noir has joined #openttd
12:13:15  *** ChanServ sets mode: +v tokai|noir
12:19:44  *** tokai has quit IRC
12:21:52  <DorpsGek> [OpenTTD/OpenGFX] glx22 updated pull request #72: Add: [Actions] Make NML warnings more visible in pull requests
12:22:39  *** glx has joined #openttd
12:22:39  *** ChanServ sets mode: +v glx
12:32:31  <DorpsGek> [OpenTTD/OpenGFX] glx22 commented on pull request #72: Add: [Actions] Make NML warnings more visible in pull requests
12:33:10  <glx> and all warnings are gone
12:33:26  <LordAro> :)
12:34:17  <glx> gimp wanting to store iCCP chunk is really annoying
12:35:18  <LordAro> 30kB gone from terrain.xcf, is all the information still there?
12:36:23  <glx> I just changed 7 pixels
12:36:44  <LordAro> very good
12:37:02  <glx> and different layers are selected as I worked on different parts
12:37:26  <glx> I don't know the xcf format but maybe that explains the difference in size
12:39:55  *** Gustavo6046 has quit IRC
12:42:05  *** Gustavo6046 has joined #openttd
12:53:53  <Eddi|zuHause> yay, finally openttd in java
12:54:36  <Eddi|zuHause> an over 10 years outdated version of openttd ported to java is what the world needed
13:07:06  <LordAro> "it is based on a version that, to my opinion had the best implementation of rail signals"
13:07:09  <LordAro> uh.
13:10:00  <glx> funny, initial import also includes all .svn subdirs
13:15:29  <LordAro> "Also, C code was easier to port to different language, than C++." i suspect this may be the true reason
13:19:30  *** WormnestAndroid has quit IRC
13:19:33  *** WormnestAndroid has joined #openttd
13:23:58  *** Fuco has joined #openttd
13:27:54  <Samu> something happened to english.txt, im getting weird conflicts
13:28:55  <Samu> i suspect this
13:29:32  *** WormnestAndroid has quit IRC
13:29:46  *** WormnestAndroid has joined #openttd
13:29:47  <glx> this commit is fine
13:30:06  *** Wuzzy has joined #openttd
13:30:30  *** Fuco has quit IRC
13:30:49  <LordAro> it does have a high chance of causing conflicts if you modify english.txt though :p
13:31:05  <glx> of course, but that's the fun part of a rebase
13:32:09  <LordAro> it also includes the string 'AIRACRAFT'
13:32:35  <Samu> what is this ###length 15 used for
13:32:45  <Samu> do i have to edit if I add strings?
13:32:47  <glx> for validation
13:32:59  <glx> size of the menu usually
13:33:22  <glx> or number of related strings
13:33:32  <Samu> how do i run this string validator locally?
13:33:41  <LordAro> on windows? with difficulty
13:34:02  <glx> it's a python script
13:34:16  <LordAro> glx: are you going to walk Samu through installing python?
13:34:20  <glx> no
13:34:29  <Samu> python for windows is a thing? :p
13:34:34  <LordAro> of course
13:35:11  <Samu> i think im changing that length to 16, i remove 1 stirng and add 2
13:35:16  <glx> windows even launch the store if you try to run python
13:35:23  <LordAro> @calc 15 - 1 + 2
13:35:23  <DorpsGek> LordAro: 16
13:35:27  <LordAro> good bet.
13:35:44  <Samu> but no way to verify, not that it matters I guess
13:35:57  <LordAro> no, OTTD itself doesn't care
13:36:05  <LordAro> (nor OTTD compile)
13:37:06  <glx> I don't really see where you are touching strings with ###length 15
13:38:19  <Samu> I removed this one STR_FRAMETIME_CAPTION_AI
13:39:12  <glx> ah ok, indeed
13:39:49  <glx> yes update the length then
13:40:03  <Samu> it's my Self-adjusting opcodes thingy
13:43:07  *** morbidbird has joined #openttd
13:46:38  <morbidbird> hi
13:48:52  <Samu> i dont understand why the order of some strings were swapped
13:49:32  <glx> it's all explained in the PR
13:50:10  <glx>
13:58:06  *** gelignite has joined #openttd
14:32:15  *** Wuzzy has quit IRC
14:33:59  *** morbidbird has quit IRC
14:34:26  *** Etua has quit IRC
14:51:31  <Samu> i wonder if my ai can get 5000 ships * 15 on a 4096x4096 map, I suspect it can, but let's test
14:52:00  <Samu> let alone 5000 rv+5000 ac
15:10:54  *** esselfe has joined #openttd
16:03:45  *** frosch123 has joined #openttd
16:14:55  <frosch123> finally there is an option for people who want to play with the original AI :)
16:16:43  *** sla_ro|master has quit IRC
16:19:27  <LordAro> lol
16:26:52  <TrueBrain> I cannot believe that is the post that summons frosch123 :P
16:27:16  <frosch123> it was very entertaining
16:27:39  <frosch123> someone woke up from cryptosleep after 16 years
16:27:39  <TrueBrain> and the first reply is just lolz
16:27:51  <TrueBrain> like, that is the point you went: nonono?
16:27:53  <LordAro> it was /interesting/ enough that i posted for the first time in 6 months
16:28:10  <TrueBrain> so this is it, I guess
16:28:11  <TrueBrain> the end
16:28:18  <TrueBrain> the one thing that wakes up all devs, is a post about Java
16:28:19  <TrueBrain> :D
16:28:46  <TrueBrain> I did weirder shit in my life, so I am not judging :)
16:28:47  <LordAro> i started the day by getting inexplicable errors from maven, tbf
16:29:03  <TrueBrain> I just received a sentry report I did not expect
16:29:17  <frosch123> TrueBrain: i never have the endurance to spend 1 months to port a few 100k lines :)
16:29:23  <frosch123> so i would never achieved that
16:29:36  <TrueBrain> I did OpenDUNE .. what can I say ..
16:29:47  <LordAro> and all of OTTD's infrastructure :p
16:30:05  <TrueBrain> that is sensible (isn't it? :D)
16:30:33  <TrueBrain> too bad he ported such an old version .. makes multiplayer a lot harder :P
16:30:36  *** virtualrandomnumber has joined #openttd
16:30:40  <LordAro> haha
16:30:41  <frosch123> but the choice "old ai" and "no yapf or anything important today" is way weirder than choosing "java".
16:30:52  <frosch123> obviously he played the original TTD, but probably never ttdp or ottd
16:31:10  <frosch123> oh right, jDrive is way closer to original TTD than to OpenTTD 12 :)
16:31:16  <TrueBrain> I would have written scripts to do the conversion, honestly :)
16:31:16  <frosch123> it's really a port of TTD, not of OTTD
16:31:56  <frosch123> <- that's the best part though :)
16:32:10  <frosch123> quite sure we still have some cases of using comma-operator in weird places :)
16:32:57  <TrueBrain> maybe he can now explain how the rv takeover code works?
16:33:19  <frosch123> no, he ported ottd 0.4.something
16:33:31  <frosch123> rv takeover is probably from ottd 0.6/0.7
16:33:38  <frosch123> with trams and articulated vehicles
16:33:46  <TrueBrain> booooo
16:34:16  <TrueBrain> I like it when people do crazy shit :D This should go on the wall :)
16:38:18  <DorpsGek> [OpenTTD/bananas-server] TrueBrain opened pull request #62: Fix: setting up a websocket connection causes an exception
16:38:40  <TrueBrain> found the reason sentry complained :)
16:42:09  <DorpsGek> [OpenTTD/bananas-server] TrueBrain merged pull request #62: Fix: setting up a websocket connection causes an exception
16:42:33  <DorpsGek> [OpenTTD/bananas-server] TrueBrain created new tag: 1.4.2
16:46:43  <LordAro> TrueBrain: did you get an exception when andy tried to upload a (ogfx) zip ?
16:46:55  <TrueBrain> a few days ago?
16:46:57  <LordAro> yeah
16:47:03  <TrueBrain> you digging up old news? :P
16:47:12  <LordAro> possibly
16:47:12  <TrueBrain> tldr: wasn't my fault
16:47:18  <TrueBrain> only important part :)
16:47:19  <LordAro> i didn't see anything about it being fixed
16:47:21  <LordAro> :p
16:47:36  <TrueBrain> I have to fix it some day, but the fix will be to refuse it without the 500
16:47:39  <TrueBrain> it was still invalid :)
16:47:48  <LordAro> hehe
16:47:57  <TrueBrain> glx is doing the real work :)
16:48:15  <TrueBrain> guess 7.1 is around the corner? :)
16:48:37  <TrueBrain> (in case you missed it, OpenGFX generated an obg that was an invalid ini-file)
16:48:52  <TrueBrain> because some translator decided to add a case-variant
16:49:02  <TrueBrain> no clue why or how that is possible .. but that happened :)
16:49:09  <LordAro> oh, was that the cause of the error? rather than because it was a zip
16:49:22  <TrueBrain> that was causing the 500
16:49:25  <TrueBrain> no clue about zip etc
16:49:53  <TrueBrain> maybe frosch123 knows, but this is weird:
16:50:16  <glx> openttd doesn't care it's not valid, because it parses key=string and just ignore the broken line
16:51:06  <glx> TrueBrain: the translator did it on sfx and msx too
16:51:30  <glx> guess they don't understand the use of cases
16:51:46  <TrueBrain> on the other hand, why do we allow it? :P
16:51:52  <LordAro> should be rejected by eints, presumably
16:51:59  <TrueBrain> anyway, bananas-server fixed ... Content Service via wasm works again :)
16:53:49  <glx> cases are supported by newgrf, and eints works fine, it's just ob[gms] for basesets that are a special case
16:53:56  <frosch123> TrueBrain: it's a comment from the translator :p it seems to mean "status: omitted"
16:54:03  <frosch123> some copy/paste mistake or so
16:54:05  <TrueBrain> there are 2 cases
16:54:37  *** Progman has joined #openttd
16:54:38  <glx> and
16:55:22  *** Progman has quit IRC
16:55:40  <TrueBrain> <- new snapshot .. still looking good :)
16:56:16  *** Progman has joined #openttd
16:56:37  <frosch123> i am more concerned about the reordering in
16:58:09  <LordAro> TrueBrain: what happened 08/22-08/27 ? strangely high amount of timeouts?
16:58:28  <glx> bugs were not fixed yet ?
16:58:53  <peter1138> How do I play this java version then? It runs everywhere apparently...
16:59:02  <peter1138> (Also the newgrf support will be lolz)
17:00:43  <frosch123> i am not sure the porter is aware of that
17:00:53  <frosch123> they mentioned they want to add bananas support
17:00:58  <frosch123> but literally nothing on bananas would work
17:01:02  <frosch123> not even basesets
17:01:17  <LordAro> why wouldn't it?
17:01:26  <frosch123> basesets are ottd 0.7
17:01:31  <glx> GS and AI are out too
17:01:41  <frosch123> ottd 0.4 has single hard-coded .grf files instead of action5
17:01:49  <LordAro> oho.
17:01:57  <frosch123> heightmaps are the only thing that works
17:02:56  <glx> scenarios may work too (if old enough) but the savegame format will be different as I understant the goal
17:03:13  <frosch123> and if they wanted to add all the newgrf stuff, they would have ported a newer version
17:03:37  <frosch123> so this "add bananas support" thing makes no sense to me
17:05:48  <LordAro> when was newgrfv8 ? (that was the one with container v2, right?)
17:06:17  <frosch123> grf8 and container2 are unrelated, but both happened with 1.2
17:06:44  <frosch123> really, they picked a version with about no newgrf support
17:07:12  <frosch123> if they had picked a revision a few months earlier, that would be literally nothing
17:07:14  <frosch123> it just started
17:07:33  <frosch123> it's really TTD, not OTTD
17:08:14  <glx> no map accessors either I think
17:08:19  <frosch123> oh, it's even before tgp
17:08:38  <frosch123> no "ottd feature" i can think of is present
17:08:45  <glx> so yeah the readme can complain about bitbanging
17:09:18  <frosch123> so really, this thing only makes sense, if you have not played ottd in 16 years
17:09:25  <LordAro> "barely beyond ludde's decompiled version" wouldn't be putting it too strongly?
17:09:39  <frosch123> otherwise you would immediately notice the lack of build-on-slopes and similar things
17:09:57  <frosch123> LordAro: 1.5 years of development
17:10:10  <peter1138> Best signals though right?
17:10:35  <frosch123> oh right, that's the solution to the mistery
17:10:46  <LordAro> "As a side bonus - effortless portability to major platforms."
17:10:48  <frosch123> they only found ottd after block signals were hidden from the gui
17:11:02  <LordAro> compared with the lack of support for all the major platforms we have now
17:11:08  <frosch123> maybe they think pbs replaced block signals entirely
17:12:11  <frosch123> LordAro: i would think you can compile java to wasm just like any other language?
17:13:19  <LordAro> oh, the guy's Russian
17:13:22  <LordAro> suddenly makes a bit more sense
17:13:33  <LordAro> also, TTD resources :o
17:14:11  <frosch123> lol, xussr.grf is there?
17:14:20  <frosch123> how is that going to work?
17:15:11  <peter1138> Well it's a very old NewGRF, it might be old enough...
17:15:47  <frosch123> maybe there were ealier versions, but the xussr i know has like 32k engines :p
17:15:59  <glx> pff the kid on discord is sometimes annoying
17:16:03  <frosch123> certainly more than the 100 supporter by ottd < 0.7
17:16:24  <peter1138> It was started in 2005.
17:16:51  <peter1138> glx, I like how vague that is.
17:17:06  <LordAro> isn't everyone on discord a child?
17:17:47  <frosch123> is there some tower defense game named "children vs boomers"?
17:17:54  *** _aD has joined #openttd
17:19:12  *** Wormnest has joined #openttd
17:22:55  <peter1138> Oh right, it has the old PBS implementation.
17:22:59  <peter1138> Hahaha
17:23:18  <frosch123> wait, is that the superior signalling?
17:24:35  <peter1138> If you like your network to deadlock or crash, yeah.
17:24:53  <LordAro> haha
17:25:36  <glx> one of the rare change we removed at one point
17:25:59  <LordAro> we removed OPF not that long ago!
17:27:48  <LordAro> oh, and he's MIT licenced it
17:27:56  <LordAro> that's a big nono
17:29:16  <frosch123> is it? it does not really share code
17:29:36  <LordAro> it's very clearly derived from OTTD
17:29:41  <LordAro> that means it must be GPL, no?
17:29:43  <frosch123> it's reverse-engineered in a clean room
17:29:51  <frosch123> just like ludde did
17:30:13  <LordAro> well this is certainly OTTD
17:30:14  <glx> first import contained a full svn checkout
17:30:16  <LordAro> and is missing the licecne
17:30:29  <glx> including .svn subdirs
17:31:39  <LordAro> makes you wonder where they got it from, given the svn servers have been turned off for 3 years now
17:31:56  <frosch123> LordAro: <- it's there
17:32:12  <LordAro> hmmm
17:32:13  <frosch123> readme says ottd 0.4.1
17:32:32  <frosch123> but it may be some intermediate revision, just before yapf started
17:35:00  <TrueBrain> LordAro: the timeouts in that period were mostly a bug I fixed, and fixing that meant "timeout" split between timeout and aborted :)
17:35:11  <LordAro> :)
17:35:13  <TrueBrain> it is also why there are more TURN connections in that time
17:35:38  <TrueBrain> from about 08/30 I haven't really changed anything anymore :)
17:36:58  <LordAro> worth posting?
17:37:51  *** morbidbird has joined #openttd
17:39:13  <TrueBrain> it would help if that license part was not in there, it kinda makes the last part a bit weird
17:39:27  <frosch123> LordAro: there is probably some partial build-on-slopes
17:39:31  <TrueBrain> but it is an important thing to mention :)
17:39:47  <TrueBrain> guess I can steal your thunder and post that :P
17:39:55  <LordAro> go ahead :)
17:40:06  <peter1138> The license is the only important issue
17:40:11  <peter1138> The rest is up to them to worry about.
17:40:15  *** jottyfan has joined #openttd
17:40:37  <LordAro> it is
17:40:46  <frosch123> glx: "svn+ssh://"
17:40:47  <LordAro> just "We do think that it will simplify development of game and let us implement things that nearly impossible or very hard to implement in traditional OpenTTD architecture." rubs me the wrong way a bit
17:40:59  <_dp_> lol, NextTTD
17:41:00  <frosch123> maybe there was some svn mirror in russia somewhen?
17:41:11  <_dp_> though not that wrong I guess if you don't read it as NextOpenTTD
17:41:12  <frosch123> maybe they have the first 972 revisions there?
17:41:39  <frosch123> LordAro: it's named "jdrive"
17:42:01  <LordAro> "We're looking for a brave programmers to join a new project: NextTTD - clone of OpenTTD rewritten in Java."
17:42:13  <glx> the window title is jTTD
17:42:27  <frosch123> LordAro: oh right, maybe point them to openlomo?
17:42:34  <LordAro> haha
17:45:10  <frosch123> so, apparently it's based on r3352 from 2005-12-28
17:47:10  <frosch123> a few revisions later there is a fix for FS#36 :)
17:47:37  *** jottyfan has quit IRC
17:47:39  <frosch123> i don't see any imminent change to c++ there though, so it looks like a completely random revision they found somewhere
17:48:13  <TrueBrain> LordAro: done
17:50:55  <TrueBrain> removing the copyrighted files is an interesting problem :P
17:51:03  <TrueBrain> rewriting git history? :D
17:51:27  <LordAro> ORCT2 didn't care too much
17:56:09  <TrueBrain> I just love how he is this optimistic changing languages is going to change anything :D
17:56:11  <TrueBrain> time will tell ;)
17:56:20  <Samu> i wonder if title screen competition will even get an entry
17:56:46  <Samu> it reads to me as overly complicated
17:57:15  <frosch123> yeah, it sounds similar as newmap and 32bpp stuff :)
17:57:25  <frosch123> 32bpp will add trains smooth turning
17:57:34  <TrueBrain> I have never attempted to rewrite very file in a more readable way .. NEVAH
17:57:43  <TrueBrain> very = every
17:59:00  <frosch123> <- is it more readable ?
17:59:25  <TrueBrain> dunno, I don't speak Java :D :D
17:59:39  <frosch123> <- hey, your AI is the default
17:59:56  <TrueBrain> \o/
18:00:16  <TrueBrain> I forgot how hacky the old AI was
18:00:21  <TrueBrain> I cannot believe CS managed to write it
18:00:25  <TrueBrain> so many places to hook into the code
18:01:27  <frosch123> <- letter from truelight
18:01:44  <TrueBrain> those times we signed our name in long comments
18:01:44  <TrueBrain> so cute
18:01:48  <frosch123> less than 3 months old at that revision
18:02:10  <frosch123> (the comment, maybe truelight as well)
18:02:53  <TrueBrain> I like that even in 2005 I was critiquing my own code without issue :P
18:04:22  <frosch123> aw, it's funny how close java is to C
18:04:41  <TrueBrain> your and mine definition of funny widely differ ;)
18:04:47  <frosch123> c code that would be considered awful today is still valid java
18:04:58  <frosch123> they did not attempt to make that stuff invalid from the start
18:05:58  <TrueBrain> so C++ ? :D
18:06:25  *** Flygon has quit IRC
18:06:30  <frosch123> sometimes i think java is closer to c than c++ is to c
18:06:46  <TrueBrain> :D
18:19:01  <morbidbird> recently i did a lot of work in java
18:19:10  <morbidbird> i wish i had ended up with java as my "main" language at age 13 instead of c++
18:19:17  <morbidbird> would have been so productive. the IDE codes for you, no memory issues
18:19:35  <morbidbird> instead of being like 14, writing big programs in C++ and not knowing what a debugger is
18:19:36  <morbidbird> shotgun debugging
18:19:40  <frosch123> you can still switch to cobol :)
18:19:51  <morbidbird> yes things were better then
18:20:13  <morbidbird> on that topic, would you accept a c++ AI pathfinder in openttd
18:20:18  <morbidbird> i think there was discussions about it somewhere
18:20:31  <frosch123> it was rejected at least 3 times before :)
18:20:49  <morbidbird> because it might run too long and it needs to do callbacks? those are the main issues?
18:21:07  <morbidbird> you're just anti-robots
18:24:31  <LordAro> not really "main issues", more "it was decided they were a bad idea 15 years ago and all support for them ripped out
18:24:34  <LordAro> "
18:24:45  <LordAro> "also we have a proper scripting framework for them now"
18:29:30  *** sla_ro|master has joined #openttd
18:31:07  <FLHerne> morbidbird: One objection was that it would effectively force AIs to use the same pathfinding strategy and limit innovation there
18:31:56  <FLHerne> (because few AI developers would choose to write their own and inevitably slower pathfinder given a built-in one)
18:32:19  <morbidbird> aha
18:32:33  <FLHerne> morbidbird: My vague thoughts were that pathfinding could be sped up with more generic functions
18:32:40  <FLHerne> especially at a high level
18:32:54  <morbidbird> maybe there are ways to get around the relative slowness, like cutting the path into sections
18:33:15  <FLHerne> things like "what's the lowest point in this region" or "is there any obstacle on this [line/list of tiles]"
18:33:55  <LordAro> FLHerne: of course "few AI developers would choose to write their own" is true anyway
18:33:59  <FLHerne> That, or add a built-in pathfinder with some sort of really configurable cost function almost like a SQL query
18:34:13  <FLHerne> yeah, they all use SuperLib or whatever the other one is
18:35:04  <LordAro> i don't think even superlib has its own pathfinder, they all use Pathfinder.Road/Rail
18:35:05  <FLHerne> but I think there are a lot of functions that could allow AIs to find nice paths quicker without literally being a "get me from A to B" function
18:35:49  <FLHerne> My *other* idea was that similar functions could allow doing map generation in Squirrel
18:36:11  <FLHerne> I assume someone else has had that
18:36:16  <LordAro> scriptable mapgen has definitely been floated before
18:36:23  <LordAro> scriptable anything has almost certainly been floated before
18:36:29  <FLHerne> Yeah
18:36:39  <FLHerne> It's kind of a pity OTTD uses Squirrel IMO
18:37:08  <FLHerne> it's not particularly well-suited to most of these things, at least performantly
18:37:14  <LordAro> mm
18:37:17  <Timberwolf> I did write one which used a relaxed A* algorithm to speed up routefinding, although it was pretty closely based on the Pathfinder.Rail library. And also I never released it.
18:37:23  <LordAro> trouble is we're stuck with it now
18:37:31  <LordAro> unless someone wants to write a squirrel parser in lua or something
18:37:48  <Timberwolf> I also tried writing a double-track pathfinder to build in a more "human-like" style, but that turns out to have a *lot* of corner cases.
18:37:56  <LordAro> trAIns does that
18:38:02  <LordAro> only one to do so, afaik
18:38:09  <FLHerne> ...and runs into a lot of the corner cases :p
18:38:11  <LordAro> yup
18:38:21  <LordAro> impressive on flat maps though ^^
18:38:39  <FLHerne> There's that one that builds double-track grid networks
18:39:23  <LordAro> choochoo does well, yeah
18:39:48  <Timberwolf> I have wondered about dividing the map into grids for doing more "human-like" networks, and maintaining a list of whether you already have track that connects to each edge. Probably a nightmare to get working well, though.
18:39:57  <FLHerne> Oh, and RailwAI
18:40:17  <FLHerne> yeah, ChooChoo was the grid one I was thinking of
18:40:30  <FLHerne> IMO it gets confused a lt
18:40:36  <Timberwolf> AIs are kind of bad at debugging problems with a big connected networks.
18:40:36  <FLHerne> but the networks look nice
18:40:39  <FLHerne> *IME
18:40:59  <Timberwolf> It's hard to encode the kind of, "that looks wrong" followed by "OK, why are they doing that" which comes naturally to a player.
18:44:08  *** Gustavo6046 has quit IRC
18:46:02  *** Gustavo6046 has joined #openttd
18:47:56  <DorpsGek> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master
18:47:57  <DorpsGek>   - Update: Translations from eints (by translators)
19:00:05  <_aD> FLHerne: but does not travel past the corner cases because I'm not an animal, I have 90° turns disabled.
19:01:00  <FLHerne> heh
19:04:23  <FLHerne> hm, trAIns has its own paper
19:06:43  <morbidbird> the way to do double track is to reduce the cost of moving next to another track
19:06:58  <frosch123> yes, that paper was important to make ottd noteworthy for wikipedia :)
19:07:11  <morbidbird> but the actual "ai" wasn't that impressive was it?
19:07:18  <morbidbird> it was just tweaking parameters?
19:07:52  <morbidbird> seeing a screenshot of choochoo is what got me into openttd ai :P
19:14:53  *** Wolf01 has joined #openttd
19:40:46  *** virtualrandomnumber has quit IRC
19:40:53  *** Gustavo6046_ has joined #openttd
19:41:45  *** WormnestAndroid has quit IRC
19:44:58  *** Gustavo6046 has quit IRC
19:44:58  *** Gustavo6046_ is now known as Gustavo6046
19:45:20  *** WormnestAndroid has joined #openttd
19:48:34  *** iSoSyS has joined #openttd
19:50:23  *** iSoSyS has quit IRC
19:53:23  *** WormnestAndroid has quit IRC
19:53:37  *** WormnestAndroid has joined #openttd
20:02:04  <Samu> trAIns is impressive
20:02:57  <FLHerne> Yeah
20:03:26  <FLHerne> The track pathfinder isn't especially complicated, but it's an elegant approach and seems to be tuned well
20:04:17  <FLHerne> The junctions, stations and network design are quite complex and produce an impressive result
20:10:52  *** Gustavo6046_ has joined #openttd
20:11:10  *** Gustavo6046 has quit IRC
20:11:10  *** Gustavo6046_ is now known as Gustavo6046
20:20:41  *** Strom has quit IRC
20:22:23  *** Strom has joined #openttd
20:53:00  *** frosch123 has quit IRC
21:01:54  *** Wolf01 has quit IRC
21:02:41  *** morbidbird has quit IRC
21:16:28  *** esselfe has quit IRC
21:20:26  *** esselfe has joined #openttd
21:20:34  <_aD> Have I managed to press a magic button that causes vehicle windows to keep opening towards the lower-right of my screen?
21:20:46  *** nielsm has quit IRC
21:21:32  <_aD> I've tried to set and reset the vehicle window size...
21:21:54  <glx> maybe a hidden CTRL feature
21:22:16  * _aD is somewhat out of CTRL
21:23:45  <_aD> ah, it relates to the positions of other on-screen windows.
21:25:33  *** HerzogDeXtEr has joined #openttd
21:27:28  *** gelignite has quit IRC
21:28:34  *** Samu has quit IRC
21:32:19  *** sla_ro|master has quit IRC
21:54:54  *** Strom has quit IRC
21:56:34  *** Strom has joined #openttd
22:04:05  *** roadt__ has quit IRC
22:05:30  *** WormnestAndroid has quit IRC
22:06:27  *** WormnestAndroid has joined #openttd
22:47:23  *** Progman has quit IRC
23:32:33  *** Gustavo6046_ has joined #openttd
23:35:46  *** Gustavo6046 has quit IRC
23:35:51  *** Gustavo6046_ is now known as Gustavo6046

Powered by YARRSTE version: svn-trunk