IRC logs for #openttd on OFTC at 2020-05-06
⏴ go to previous day
00:06:50 <glx> or it's broken, it's very hard to dig in nml source
00:09:34 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh opened issue #8116: Old patchpack savegame versions will soon overlap https://git.io/JfZsT
00:11:27 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8116: Old patchpack savegame versions will soon overlap https://git.io/JfZsT
00:46:08 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison commented on issue #8116: Old patchpack savegame versions will soon overlap https://git.io/JfZsT
01:17:24 *** WormnestAndroid has quit IRC
01:18:37 *** WormnestAndroid has joined #openttd
01:22:49 *** ChanServ sets mode: +o Yexo
02:07:38 *** ChanServ sets mode: +o Yexo
02:39:57 *** WormnestAndroid has quit IRC
02:41:40 *** WormnestAndroid has joined #openttd
02:46:48 *** WormnestAndroid has quit IRC
02:48:37 *** WormnestAndroid has joined #openttd
02:51:07 *** WormnestAndroid has joined #openttd
03:12:27 *** WormnestAndroid has quit IRC
03:12:44 *** WormnestAndroid has joined #openttd
03:17:19 *** ChanServ sets mode: +v tokai
03:25:38 *** ChanServ sets mode: +o Yexo
03:43:24 *** gelignite_ has joined #openttd
03:54:55 <DorpsGek_III> [OpenTTD/OpenTTD] Eddi-z commented on issue #8116: Old patchpack savegame versions will soon overlap https://git.io/JfZsT
07:12:13 *** snail_UES_ is now known as Guest24254
07:12:13 *** snail_UES_ has joined #openttd
08:34:56 *** sla_ro|master has joined #openttd
08:45:55 *** andythenorth has joined #openttd
10:22:53 *** Smedles has joined #openttd
10:44:38 *** supermop_Home_ has quit IRC
10:53:22 <LordAro> guess we should release 1.10.2, huh?
10:53:51 <_dp_> I don't think it will help
10:54:23 <_dp_> As far as I understand that vehicle desync that's not it
10:54:46 <LordAro> well we need another set of debug save files then
10:54:57 <LordAro> (assuming it's not whatever changes you've made, of course)
10:55:41 <_dp_> it's similar to reddit logs, players get desynced immediately after joining
10:56:32 <LordAro> have you tried reproducing with the server save? you make it sound like it'd be easy to reproduce..
10:57:19 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro approved pull request #8111: Fix: reset roadtype/streetcartype info for non-road bridges https://git.io/JfZgE
11:02:40 <_dp_> LordAro, ehm, what save? at's also pretty random as it more or less worked for a month and then this happened
11:03:03 <_dp_> and ofc it all happened when I was alseep and server already restarted on schedule :/
11:03:16 <LordAro> _dp_: as in, the save that the server was running
11:09:52 <_dp_> it's one minute before sabi desyncs but that may've been vehicle desync
11:11:05 <_dp_> instadesyncs start like 2 hours later
11:14:25 <LordAro> really need a proper reproducer to do anything about it
11:14:29 <_dp_> also not everyone desyncs, some players are fine while some can't connect
11:14:35 <LordAro> until then, it could just be a broken client and/or server
11:16:56 <_dp_> I'll try to at least save command log next time
11:18:25 <_dp_> also chris should be able to tell how it is from player pov once he wakes up
11:18:46 <_dp_> coz afaik so far there are 0 reports from players somehow, even on reddit %)
12:19:14 *** gelignite has joined #openttd
12:46:55 *** ChanServ sets mode: +o Yexo
12:53:29 <Samu> that legend didn't fit :(
12:54:41 <Samu> priority queue 2 hasn't finished yet
12:54:52 <Samu> will ruin the graph once it finishes anyway
12:59:09 <Samu> ticks needed to road pathfind from the center of 9 towns to the center of the others, looping all towns
12:59:18 <Yexo> To me this looks like an argument *against* a native priority queue. A gain of about 20% over an optimized squirrel implementation isn't very much, and the more C++ objects there are the more difficult fixing things like #6322 will become
12:59:58 <Yexo> tracking memory used by squirrel/AI
13:00:40 <Yexo> Samu: have you experimented with the estimate function yet? If not, try returning 1.2 times whatever the estimate function returns by default.
13:01:13 <Yexo> The pathfinder will find paths much quicker, but it will find slightly less optimal paths (up to 1.2 times the cost)
13:05:06 <Samu> im using an estimate of 1
13:05:23 <Samu> the original road pathfinder doesn't have this estimate, so it's 1
13:06:07 <Samu> i experimented it on my canal pathfinder, it's great! it's like magic
13:09:46 <Samu> PR#8091 is the real winner
13:10:38 <Samu> the AIList Sorted List based ones look like they perform identical to it, but in reality, it devours memory :(
13:11:14 *** andythenorth has left #openttd
13:24:24 <Samu> terron priority queue is surprisingly better than I expected
13:24:36 <Samu> i have to investigate its workings better
13:37:24 *** WormnestAndroid has quit IRC
13:37:40 *** WormnestAndroid has joined #openttd
14:09:09 *** andythenorth has joined #openttd
14:09:15 <andythenorth> someone ask peter if it's lunch pls
14:48:59 <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #118: Fix: Allow calculation for palette in spritelayout https://git.io/JfZi5
14:52:19 *** mcbanhas has joined #openttd
14:58:06 *** snail_UES_ has joined #openttd
15:04:24 <Samu> i can gain a little few more ticks by using _count
15:04:36 <Samu> 12780 in the graph becomes 12768
15:07:00 <FLHerne> mcbanhas: > If referring to the player, always try to use sentence structures that address players as "you". < doesn't seem relevant under "Bias-free communication"
15:08:32 <FLHerne> Some of the other suggestions don't apply to any texts that exist in OpenTTD, but future-proofing can't hurt...
15:09:04 <mcbanhas> Yeah, I'm future proofing it, pretty much
15:09:28 <mcbanhas> Mainly because at some point, I would like to propose and write an in-game manual.
15:09:39 <FLHerne> mcbanhas: The screenshot for "Descriptive texts" violates its own guideline re. ending in a period ;-)
15:10:52 <mcbanhas> Last sentence of tooltips are an exception. You can see it on the formatting tooltips section.
15:11:03 <mcbanhas> Although I should reword that, thanks.
15:11:07 <FLHerne> (also, I don't think the text is optimal, but that's a different thing)
15:11:20 <FLHerne> Ok, then I think that's a mistake
15:11:30 <FLHerne> At least for longer tooltips
15:11:45 <mcbanhas> Not ending in a period?
15:11:52 *** supermop_Home has joined #openttd
15:12:40 <mcbanhas> Most tooltips in game have that. The last sentence almost never ends in a period, so I just made a rule out of it.
15:12:49 <FLHerne> Specifically, that with multiple sentences the preceding ones have periods, but not the last one
15:13:09 <mcbanhas> It's easily fixable though.
15:13:34 <FLHerne> I think ignoring periods for single-sentence tooltips is fine, and maybe better
15:14:11 <FLHerne> But having periods on some but not all sentences in a text is definitely wrong
15:14:30 <mcbanhas> That's not a bad idea, because then we could distinguish between title-only tooltips and description only tooltips.
15:15:03 <mcbanhas> Well, it's not wrong if you make a rule out of it XD
15:15:23 <FLHerne> It's wrong in that I find it disconcerting to read :P
15:16:08 <mcbanhas> It's totally fixable though. Although I will have to redo most tooltips screenshot examples all over ahaha
15:16:12 <FLHerne> Pedantically, "FAQ" is an initialism and not an acronym. No-one cares though.
15:17:32 <DorpsGek_III> [OpenTTD/OpenTTD] ldpl commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
15:17:52 <FLHerne> Why is BaNaNaS called that? It's a terrible name :-/
15:18:29 <mcbanhas> It's not up to me to question that. xD
15:18:55 <FLHerne> "Base graphics/sound And Newgrfs And Noais And Scenarios."
15:19:06 <FLHerne> Even the explanation has bizarre and inconsistent casing
15:19:16 <FLHerne> (both within itself, and with the acronym)
15:20:20 <FLHerne> I disagree with the "Extended Generic Road Vehicle and Tram Set or eGRVTS is a newgrf vehicle set by Zephyris."
15:20:55 <FLHerne> I doubt a majority of users even know the full name
15:21:05 <mcbanhas> it's just a dummy example man
15:21:09 <FLHerne> "Extended Generic Road Vehicle and Tram Set"
15:21:31 <mcbanhas> I took that one from the wiki actually:
15:22:18 <FLHerne> Ok, but I think the principle is wrong
15:22:32 <FLHerne> Yes, I know what it is
15:22:53 <mcbanhas> It's an example, if you can think of a better one, let me know
15:23:21 <FLHerne> Making a sentence twice as long, and including a full name that no-one actually uses, isn't worth it just to avoid a lowercase first letter
15:23:54 <FLHerne> To be fair, I see someone called the wiki article that
15:25:02 <FLHerne> "...on the link down below." <- redundant, just "below"?
15:25:33 <FLHerne> There's an inconsistency between examples
15:25:36 <FLHerne> > The Dash is a diesel multiple unit (DMU) available since 1984
15:25:47 <FLHerne> > The EMU (electric multiple unit) is a type of train
15:26:11 <FLHerne> Should the acronym or the spelled-out version go first?
15:26:24 <mcbanhas> Yeah, I forgot to fix that one, thanks
15:26:27 <FLHerne> Possibly there should be a rule for that, it's a fairly common case
15:26:41 <FLHerne> (unless there is one and I've not got to it)
15:26:50 <FLHerne> I prefer the first one, incidentally
15:28:15 <Samu> how much code magic is in this sentence? return !_count-- ? null : _queue.pop()[0];
15:29:15 <glx> speed would be the same with parenthesis
15:29:40 <Samu> my main issue is what if :count becomes -1
15:29:56 <Samu> next time it checks, it tries to _queue.pop() instead :|
15:30:22 <FLHerne> Samu: Why not just reverse it?
15:30:42 <FLHerne> `return _count-- ? _queue.pop()[0] : null
15:31:51 <FLHerne> I guess it could overflow, eventually
15:31:55 *** arikover has joined #openttd
15:33:39 <FLHerne> mcbanhas: There's also an inconsistency in the tooltip examples
15:34:17 <FLHerne> "Clear all objects" vs "Place%bs%b a railway bridge"
15:34:28 <FLHerne> Nope, my formatting is broken
15:34:43 <FLHerne> But anyway, pick a tense
15:35:48 <mcbanhas> I had that one noted already, but thanks.
15:36:05 <FLHerne> "Clear ...", "Place ..." probably better?
15:36:18 <mcbanhas> I set a rule for that I think
15:36:31 <mcbanhas> Check the writing section for tooltips.
15:37:05 <mcbanhas> But yeah, preferably third person minus pronoun
15:38:53 <FLHerne> I hadn't got to that yet ;-)
15:40:08 <FLHerne> "March the 3rd" is a can of worms about date formatting :P
15:40:47 <mcbanhas> It's the current formatting
15:40:56 <mcbanhas> Well technically is 3rd of March
15:41:12 <mcbanhas> so I'll amend that example
15:41:35 <LordAro> i don't think it should have "of" or "the" in it
15:41:46 <mcbanhas> I'll fix the example
15:42:07 <FLHerne> In-game date is just "3rd Mar 2020"
15:42:36 <FLHerne> Which should probably have a period, given it's an abbreviation...
15:43:07 <FLHerne> I didn't see anything explicitly about preferred spelling
15:43:29 <mcbanhas> It would look a bit jarring though. I would be a good exception.
15:43:40 <mcbanhas> I haven't set rules for abbreviations, tbh, just acronyms
15:43:52 <FLHerne> The MoS itself uses US/Oxford spelling, so I guess that implies it
15:45:23 <mcbanhas> The manual is intended to cover both US/UK variants, sorta like the M$ Style Guide, but I can include a section for regional spelling, yes. :)
15:46:11 <mcbanhas> Although it would be little more than a reference to a couple of major dictionaries.
15:47:23 <glx> US is managed by translators
15:48:24 <mcbanhas> In an ideal world we would have a style guide for EVERY language xD, or least style notes.
15:49:09 <mcbanhas> This is actually a widespread problem in the free software community
15:49:25 <Samu> 12791 ticks, if i get a result lower than this, then _queue.len() ends up being faster overall
15:49:37 <FLHerne> mcbanhas: Specifically, in the UK you can have either "recognise" or "recognize", and it looks odd if they're inconsistent
15:49:48 * FLHerne tries to find any examples currently used in-game
15:50:10 <mcbanhas> There are no unifying textual guidelines or manuals of style for Linux
15:50:27 <mcbanhas> It's a mess, someone should do something about it
15:51:36 <mcbanhas> M$ itself only has a general style guide for English, and then style notes for every other language.
15:51:54 <mcbanhas> Which is an elegant solution, although not perfect
15:52:28 <Samu> 12837, wow _queue.len() disappointed
15:52:36 <FLHerne> In-game uses "penalises"
15:53:45 <mcbanhas> I'd frankly prefer using a Z
15:54:43 <FLHerne> Half a stale bread roll just fell down my chimney
15:54:43 <mcbanhas> Tbh, everything is preferable to the current state of in-game text.
15:55:10 <glx> french translation in windows is sometimes very bad
15:55:23 <FLHerne> I guess a seagull must have dropped its lunch
15:55:27 <glx> and it's worse since win10
15:56:11 <LordAro> FLHerne: or someone is throwing bread at you
15:56:31 <LordAro> purely on the ise vs ize front, english.txt is internally consistent as best as i can tell
15:56:41 <mcbanhas> glx, that's surprising to hear. The French academy goes to great lengths to ensure they don't rely on neologisms and keep all the software lingo up to date. Apparently the translators don't seem to follow then.
15:56:53 <LordAro> ( grep -o "\b[A-Za-z]*i[zs]e[A-Za-z]*\b" src/lang/english.txt | sort )
15:57:03 <FLHerne> They'd have to throw it pretty accurately to get it down my chimney
15:57:26 *** Wormnest has joined #openttd
15:58:15 <mcbanhas> "Half a stale bread roll just fell down my chimney" <- This is the most British thing I've read in a while xD
15:58:50 <glx> mcbanhas: for some time "update history" in windows update window was translated "mettre à jour l'historique" while it should be "historique des mises à jour"
15:58:54 <FLHerne> LordAro: Except that it has both "authorized" and "unauthorised" :P
15:59:14 <LordAro> FLHerne: ah right, yes
15:59:37 <FLHerne> LordAro: Adding `| uniq` to your command helps
16:00:43 <mcbanhas> glx, ouch, that's really bad
16:01:32 <glx> and it was correct before win10 (now it's fixed btw), but it's weird they could mess like that
16:02:33 <mcbanhas> Facebook also messes up translations quite often
16:03:21 <milek7> aliexpress is ridiculously bad ;P
16:04:24 * mcbanhas used to work for a Facebook localization contractor
16:05:14 <Samu> 12768 with overflow, or 12791 with code correctness
16:05:40 <glx> I know for most games the translators only receive the strings without context, but for an OS it should be different
16:06:44 <nielsm> hm, MicroProse is back as a game publisher, with some of the original people at the top... I wonder if that will mean anything for the TT copyrights and trademarks
16:06:46 <mcbanhas> Yeah, but localization always takes a low priority in software development
16:07:31 <mcbanhas> Well, Chris Sawyer Games owns TT now I think
16:07:48 <glx> chris sawyer agency handles them yes
16:07:50 <nielsm> ah yeah the 2013 mobile game definitely used the name :)
16:08:07 <mcbanhas> X-Com is owned by 2K
16:08:44 <mcbanhas> Microprose possibly still has the rights for all the flight simulators
16:09:32 <mcbanhas> nielsm, can you link me to news about the Microprose return?
16:09:54 <nielsm> just saw the headline on slashdot
16:10:13 <nielsm> (I haven't clicked through to slashdot in don't-know-how-long, just reading headlines via RSS)
16:10:41 <Yexo> mcbanhas: you give an example on how to spell out acronyms ("diesel multiple unit (DMU)") but then under "General capitalization rules" swap the order ("The EMU (electric multiple unit) is")
16:10:45 <mcbanhas> They're doing military stuff again. Back to the roots I guess
16:11:13 <mcbanhas> Yexo, that was already noted by FLHerne, but thanks :)
16:14:01 <mcbanhas> This is a very interesting development. They're completely returning to their roots. They probably found out the niche market can pay off, since Paradox has been making a ton of money with 4x stuff.
16:15:06 <mcbanhas> And besides they do have the expertise and a lot of former military personnel in their contacts.
16:17:35 <milek7> for what reason games still need 'publisher' now?
16:17:49 <Yexo> mcbanhas: under "UI labels": "as little words" should be "as few words" I think
16:18:39 <milek7> it's not like they need to press CDs and ship boxes around the world
16:19:21 <glx> milek7: to have money to develop
16:19:40 <glx> and handle all the paperwork
16:20:35 <mcbanhas> Yes, exactly wht glx just said
16:21:00 <mcbanhas> Publishers handle certification with online platforms for example
16:21:27 <glx> and now crowdfunding is mainly used to prove there's a demand so a publisher can enter in the loop
16:22:03 <mcbanhas> If you want to publish your game on a console, there's a lot of hoops you gotta jump through. Games need to be thoroughly tested and found to be in compliance with ESRB, PEGI and other rating authorities,
16:22:08 <mcbanhas> Publishers handle all that crap.
16:23:14 <mcbanhas> They also handle producing roles and establish development milestones, coordinate progress, suggest adding/removing features as p/ time constraints and budgets.
16:23:55 <mcbanhas> And playtesting of course. :)
16:24:11 * mcbanhas briefly worked as a producer for a publisher.
16:38:07 *** andythenorth has left #openttd
17:12:38 *** Gustavo6046 has joined #openttd
17:13:46 <Eddi|zuHause> <mcbanhas> There are no unifying textual guidelines or manuals of style for Linux <-- i smell a https://xkcd.com/927/ coming
17:15:22 <mcbanhas> That's why I disagree with the "just fork" mentality
17:16:09 <Wolf01> Eddi|zuHause: self fulfilling prophecy
17:17:06 <milek7> to be honest, I don't find it significant problem
17:17:52 <milek7> I don't even notice half-translated polish/english ui :P
17:22:46 <Samu> you already had your own github thingy apparently
17:24:27 <LordAro> "your own github thingy"
17:24:37 <Yexo> Go even further back and that code was hosted in an svn repo somehwere on openttd.org
17:25:51 <SpComb> kalliteha version 0.3.3, released in 2017
17:30:34 *** Gustavo6046 has joined #openttd
18:06:03 <FLHerne> (and wholly incompatible with any reasonable Manual of Style)
18:21:31 *** sla_ro|master has joined #openttd
18:25:26 <Wolf01> FLHerne: nice video :D
18:32:13 *** Laedek_ has joined #openttd
18:54:58 *** gelignite has joined #openttd
18:56:46 *** Progman has joined #openttd
19:16:29 *** andythenorth has joined #openttd
19:41:10 <andythenorth> is it beer time?
19:45:51 <DorpsGek_III> - Update: Translations from eints (by translators)
20:01:40 *** frosch123 has joined #openttd
20:01:56 *** Foveafox_ has joined #openttd
20:43:27 <mcbanhas> Does it have anything to do with apple's transition to Metal?
21:10:29 <Samu> I think I found a way to make my "Native Heap" aka Sorted List to not consume memory,
21:10:40 <Samu> but... it may consume cpu, have to check
21:12:05 <glx> Wolf01: yeah they marked OpenGL deprecated a long time ago
21:12:40 <TrueBrain> "not" consume memory? I call bullshit :D
21:14:01 <TrueBrain> and yet it may consume cpu
21:15:30 <Samu> gonna post the source code soon, just making sure it's working as it should first
21:17:14 <Samu> I have an "enumerator" to generate the item #id, so that i could put it as an item in the AIList
21:17:38 <Samu> the real item itself is stored in a table
21:17:48 <TrueBrain> which is memory-less? :D
21:18:03 <TrueBrain> I have a box, and that compresses ANY movie 10 times better than any known algorithm
21:18:06 <TrueBrain> it is an amazing box
21:18:42 <Samu> this enumerator starts from 0, and each time the insert is called, it is increased by 1
21:19:12 <TrueBrain> sometimes Samu , your programming powers impress me; but in reality, please buy a book on algorithms :)
21:22:10 <Samu> it's the AI 9, i called the queue SortedList
21:22:34 <Samu> it's not using much memory, and is getting closer to the performance of PR#8091
21:22:53 <TrueBrain> not much? But it wasn't using memory before :( YOU LIED TO ME :'( Jan, go back in your corner, you lied too :(
21:23:07 <DorpsGek_III> [OpenTTD/nml] frosch123 approved pull request #118: Fix: Allow calculation for palette in spritelayout https://git.io/JfZjA
21:23:09 <Samu> which could mean... PR#8091 could be closed, and not merged
21:23:34 <Samu> still making sure i can use the Exists function
21:23:35 <TrueBrain> but seriously; you are heavily reinventing wheels here. If that is your fun, continue on. But many books are written about this .. picking up one might help you improve even further ;)
21:24:49 <TrueBrain> I am sure that is mutual; nevertheless, reinventing stuff isn't really a definition of progress ;)
21:25:17 <frosch123> are you discussing using /dev/null as backup?
21:33:37 <Samu> i got the idea based of Trans AI, of fanioz
21:34:04 <Samu> he used the tile as the ID, but that sounds to me like a bad idea
21:34:55 <Samu> i reworked it, and instead of tile id, i use the enumerator to generate the unique id
21:38:23 <Samu> this way, the queue can be used for several kinds of items, not just a queue that includes a tile
21:39:38 <Samu> I dunno what to call it, so i called it _tracker
21:39:52 <Samu> maybe _enumerator would be better
21:40:18 <Samu> but it's a _count that never decreases
21:40:45 <Samu> _enumerator sounds better
21:42:55 <frosch123> well known people :)
21:49:41 *** Foveafoxy has joined #openttd
21:50:13 *** Foveafoxy is now known as Guest24308
21:58:35 <TrueBrain> you have a commit to fix frosch123 :P
21:59:44 <TrueBrain> on the other hand, that scenario only existed for a second; we could also just remove it
22:00:13 <frosch123> just implement upgrades for scenarios and heightmaps
22:00:30 <frosch123> the "age" is irrelevant, you cannot download scenarios/heightmaps for old savegames anyway
22:00:37 <TrueBrain> but that wont solve this entry ;)
22:01:08 <frosch123> didn't you assign the same unique-id to scenario upgrades in the data migration?
22:01:27 <frosch123> so, instead of the replaced-by you could also move them to the same unique-id
22:01:45 <frosch123> or that can be postponed till the api supports the upgrades
22:01:56 <frosch123> since then multiple may get migrated
22:02:46 <frosch123> anyway, i would recommend to not look too deeply into individual entires. that way leads to madness :)
22:03:02 <TrueBrain> but still .. this entry needs a resolution ;)
22:03:31 <TrueBrain> (follow the link in the PR to understand why :P)
22:04:04 <TrueBrain> this is not one of many. This is one ;)
22:04:16 <frosch123> well, did you miss the other two guys who renamed have of their chaos grfs to "zzz ignore"
22:04:26 <TrueBrain> yes, but they are not "new-games"
22:04:50 <TrueBrain> owh, some are .. they also need fixing in that case ;)
22:04:58 <TrueBrain> allowing to rename has this as side-effect
22:05:20 <frosch123> TrueBrain: as said, i would recommend to ignore them
22:05:31 <TrueBrain> I rather start with a cleanup :)
22:05:44 <frosch123> they prefer to abuse the tools they have, you can't win that
22:06:04 <TrueBrain> no, this is not abuse as such; this is not having better tools available and seeing this as only solution
22:06:07 <TrueBrain> which I fully understand
22:06:17 <TrueBrain> we need to deal with it in some matter
22:06:41 <frosch123> sure, but do that by implementing stuff in the api, not by fixing stuff manually :)
22:06:57 <TrueBrain> you don't pay my subscription :P
22:07:06 <TrueBrain> and clearly, I already wrote this PR ;)
22:07:56 <frosch123> well, i expect those PRs will take more time than migrating all authors :)
22:08:24 <TrueBrain> I mostly wonder how we can get in touch with authors .. mainly as the display-name is not migrated
22:08:29 <TrueBrain> that might be a bit challenging
22:09:06 <frosch123> yeah, no idea what to do with the displayname
22:09:46 <frosch123> i don't like to make it freely editable via a user profile. it should rather get updated everytime the package/version in edited/updated/added
22:10:12 <DorpsGek_III> [OpenTTD/OpenTTD] Yexo merged pull request #8111: Fix: reset roadtype/streetcartype info for non-road bridges https://git.io/JfG8l
22:12:06 <frosch123> originally i wondered whether we would add other oauth providers. but considering how popular gh is among the current bananas users, it's best to stay with one
22:12:28 <frosch123> i definitely did not expect 95% of people to already have a gh account
22:14:54 <Yexo> or the people without gh account are not asking for a migration (yet)
22:15:08 <frosch123> oh, i missed to mention the fun-fact. the two people who used the "z-ignore" have very similar gh/forum/ottd names, and make similar content... so i conspirated about a split personality :p
22:15:58 <frosch123> Yexo: many scenario/heightmap people were also on gh
22:17:44 <TrueBrain> frosch123: I did the same! I even double-checked if that was the case :D :D :D
22:27:50 <frosch123> hmm, should we host bananas awards? vote for the best ai/vehicle grf/heightmap/... ?
22:28:39 <frosch123> i am searching for a way to emphasise good content, that does not require a lot of moderation
22:29:47 <frosch123> that's hard to do with our public gh storage
22:30:12 <frosch123> too many commits, too much personal information, too much drama
22:31:12 <andythenorth> allow personal curators
22:31:15 <andythenorth> move the drama left
22:31:16 <frosch123> the bananas2 draft had a star-rating free for everyone, but it was private
22:31:24 <andythenorth> give people channels
22:31:33 <andythenorth> ratings are a disaster
22:32:50 <frosch123> yes, i do not want a continuous rating number. just a "this was considered good somewhen" badge
22:33:24 <frosch123> send new people to few good entries, let people dive into the mess if they want
22:34:27 <nielsm> also "this item works well with this other item"
22:35:01 <frosch123> that's not useful on bananas, you can't see that after download in the configuration guis
22:35:22 <nielsm> then make that happen!
22:35:51 <frosch123> the bananas2 draft contained a content-type "newgrf-profile"
22:39:17 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison opened issue #8117: Memory leak: All incoming admin port packets leaked https://git.io/JfnJ8
22:39:30 <andythenorth> frosch123 we should
22:47:32 <frosch123> moving stuff into afterload.cpp is always good
22:47:45 <frosch123> conversions in other files are the biggest headache
22:48:42 <glx> just now we have an improved way
22:50:16 <glx> Samu: and the unconditionnal fix was my idea ;)
22:50:43 <Samu> but I deleted the branch
22:51:22 <frosch123> why is it related to SLV_ENDING_YEAR, and not SLV_MULTITILE_DOCKS ?
22:51:22 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 opened pull request #8118: Fix #8021: limit savegame range for docking tiles fixing https://git.io/JfnUv
22:52:29 <glx> because the bug touched all savegames between both version
22:52:44 <Yexo> Nice. I was already wondering why there are so many AfterLoad* functions called unconditionally
22:52:52 <frosch123> glx: ok, can you add a comment then :)
22:54:05 <frosch123> Yexo: looks like people stopped adding stuff to InitializeWindowsAndCaches
22:54:35 <frosch123> it used to be the last and only call for stuff that is not stored and needs recomputing
23:04:59 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 updated pull request #8118: Fix #8021: limit savegame range for docking tiles fixing https://git.io/JfnUv
23:05:36 <DorpsGek_III> [OpenTTD/OpenTTD] frosch123 approved pull request #8118: Fix #8021: limit savegame range for docking tiles fixing https://git.io/JfnUo
23:06:25 <glx> I wonder if some desyncs are multitile related
23:06:55 <DorpsGek_III> [OpenTTD/OpenTTD] Yexo approved pull request #8118: Fix #8021: limit savegame range for docking tiles fixing https://git.io/JfnU1
23:07:45 <frosch123> glx: it would explain desyncs shortly after join
23:09:01 <frosch123> glx: if you have the suspicion that something is still broken, one could add recomputation of the dock areas to CheckCaches()
23:10:43 <LordAro> how would multitiles cause desyncs?
23:11:39 <glx> incorrect update in some functions causing different ship pathfinding result
23:11:39 <frosch123> the area is update incrementally during commands. but recomputed from scratch when loading a savegame
23:11:52 <frosch123> so, if the incremental updates are wrong, the new joining client desyncs
23:13:02 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 merged pull request #8118: Fix #8021: limit savegame range for docking tiles fixing https://git.io/JfnUv
23:15:07 <Yexo> I don't see how 'RemoveDockingTile' does any removing. Am I missing something or is that function name wrong/misleading?
23:17:38 <frosch123> i think a "docking tile" is a tile adjacent to a station tile, where the ship will stop
23:17:58 <Yexo> Yes, that seems to match how I understand the code now
23:18:01 <frosch123> so RemoveDockingTile is called for tiles which are no longer water
23:18:22 <Yexo> That is one tricky thing to keep correct, very easy to miss a case there
23:19:44 <frosch123> yes, sounds like adding some validation to CheckCaches would be a good thing
23:22:22 <Samu> which one is gonna be faster? _sorter.RemoveItem(ret); VERSUS _sorter.RemoveTop(1);
23:23:45 <frosch123> yep, IsPossibleDockingTile is missing MP_TREES at least
23:24:11 <frosch123> hmm, no, those tiles are not traversible
23:28:39 <Yexo> Ships can dock from within the lower part of a lock to a dock to the side
23:28:43 <Yexo> Not sure if that's intended
23:29:14 <glx> hmm there's no easy way to compare 2 TileArea
23:29:16 <frosch123> isn't GetTileTrackStatus checking for that?
23:29:38 <Yexo> frosch123: It doesn't get any direction
23:30:29 <frosch123> oh, ok, in that case it also applies to ship depots :)
23:30:49 <glx> and the docking status of the tile on map must be checked too
23:31:01 <frosch123> and docks below aqueducts :)
23:31:59 <Yexo> a dock below an aqueduct is not allowed
23:32:47 <frosch123> too bad :) so only dock next to aqueduct
23:33:03 <frosch123> quite sure some pp enabled bridges over stations
23:33:55 <frosch123> ah, no docking against the slope part, so master is save
23:34:15 <Yexo> "slope part" = sides of the dock?
23:35:07 <frosch123> IsValidDockingDirectionForDock is actually quite picky
23:35:17 <frosch123> only two tiles per dock
23:37:34 *** Foveafoxy has joined #openttd
23:38:13 *** Foveafoxy is now known as Guest24324
23:38:57 <frosch123> hmm IsValidDockingDirectionForDock is only used in some places
23:39:16 <frosch123> doesn't UpdateStationDockingTiles just ignore htose restrictions? and adds all tiles?
23:40:50 <Yexo> It calls CheckForDockingTile for all tiles, but CheckForDockingTile does contain the check
23:43:44 <frosch123> hmm, luckily freeform edges are enabled by default. otherwise i think it could leave the map on the north border
23:45:57 <Yexo> Remind me, is north top-right or top-left?
23:46:10 <frosch123> luaduck: _dp_: do people use ships on your servers, or are they banned as uncool?
23:46:40 <frosch123> both NW and NE have no MP_VOID when freeform is disabled
23:47:03 <Yexo> Where would that be a problem?
23:47:24 <frosch123> UpdateStationDockingTiles starts from a valid tile, and calls CheckForDockingTile for all neighbouring water tiles
23:47:35 <frosch123> CheckForDockingTile checks all neighbouring tiles again
23:47:59 <frosch123> so CheckForDockingTile can jump by distance 2 from the station, when the tile in between is MP_WATER
23:48:08 <frosch123> which is the case for non-freeform
23:48:45 <_dp_> frosch123, for the most part they're allowed but limited to like 25 per company
23:49:03 <_dp_> frosch123, but if you're looking for desync it happened on server with no ships at all
23:49:25 <frosch123> ok, that's what i wanted to know :)
23:51:16 <_dp_> for desync I suspect some company id or client id shenanigans
23:51:32 <_dp_> only something like that can explain all the weirdness afaict
23:52:48 <Yexo> I think I found a ship-related desync none-the-less
23:53:08 <Yexo> Building half-rail tiles next to a dock
23:54:25 <Yexo> CmdBuildSingleRail correctly keeps the docking-tile status intact, but I don't think CmdBuildRailroadTrack does
23:55:11 <Yexo> After building a diagonal rail (multiple in one click) that goes past the end of the dock, a ship refuses to dock there. Saving/loading the game resolves that, the ship can dock again
23:55:56 <luaduck> frosch123: quite often actually
23:56:14 <luaduck> often enough that we had to introduce a minimum buoys rule
23:56:26 <luaduck> to stop the pathfinder nomming the entire CPU
continue to next day ⏵
![https://i.imgur.com/JsqM3W8.png[/img]](https://i.imgur.com/JsqM3W8.png%5B/img%5D){kind=link}