IRC logs for #openttd on OFTC at 2014-07-25
⏴ go to previous day
01:27:10 *** Extrems has joined #openttd
02:04:25 *** luaduck is now known as luaduck_zzz
02:39:03 *** tokai|mdlx has joined #openttd
03:37:09 *** Hazzard_ has joined #openttd
03:55:40 *** Supercheese has joined #openttd
04:56:16 *** Eddi|zuHause has joined #openttd
05:18:58 *** Smedles_ has joined #openttd
05:24:46 *** Smedles has joined #openttd
05:28:17 *** HerzogDeXtEr has joined #openttd
06:01:45 *** Alberth has joined #openttd
06:01:45 *** ChanServ sets mode: +o Alberth
06:05:15 *** Smedles_ has joined #openttd
06:22:08 *** HerzogDeXtEr1 has joined #openttd
06:29:02 *** [1]Tuhin has joined #openttd
06:34:01 *** [1]Tuhin is now known as Tuhin
06:45:01 *** [1]Tuhin has joined #openttd
06:49:08 *** [1]Tuhin is now known as Tuhin
07:11:19 *** [1]Tuhin has joined #openttd
07:12:40 <Flygon_> Forgot savegames are in My Documents
07:12:43 <Flygon_> Meeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeerde
07:12:54 *** Flygon_ is now known as Flygon
07:14:58 <planetmaker> Flygon, the *only* important folder should be the one in My\ Documents
07:15:58 <Flygon> But OTTD's the only app I've had that actually utilizes My Docs x.x
07:16:03 *** [1]Tuhin is now known as Tuhin
07:16:51 <Flygon> I'm lucky I stuffed a fair few saves onto Dropbox
07:17:10 *** Smedles has joined #openttd
07:19:19 *** andythenorth has joined #openttd
07:19:25 <andythenorth> maybe the cargos miss an action 4 for names
07:20:52 *** [1]Tuhin has joined #openttd
07:21:09 <andythenorth> I don’t see an action 4 in the nfo
07:22:31 *** Smedles has joined #openttd
07:23:30 <andythenorth> 4 * 10267 04 00 FF E3 \wxD000
07:23:37 <andythenorth> that’s setting strings for trains?
07:24:05 <planetmaker> probably. action4, feature0, I think
07:24:19 <andythenorth> feature0 is trains
07:24:45 <andythenorth> not convinced that cargos will be able to use those strings
07:25:12 <andythenorth> although it’s a long time since I touched action4
07:25:23 *** [1]Tuhin is now known as Tuhin
07:25:31 <peter1138> that's probably a BAD FEATURE
07:26:22 <andythenorth> what has it done to offend YOU?
07:28:14 <Flygon> Okay, my savegames are hilariously outdated
07:28:43 <andythenorth> (UNDEFINED STRING) is a BAD FEATURE
07:28:48 <andythenorth> a BAD FEATURE of FIRS
07:28:54 *** Smedles_ has joined #openttd
07:29:12 <andythenorth> action4 - it doesn’t seem to matter what feature is set
07:29:28 <andythenorth> 00 (trains) or (0B) cargos still cause some texts to show up in industries
07:29:31 <andythenorth> but not all of them :P
07:31:01 <andythenorth> prop 09 is \wxDC00 for all cargos
07:31:09 <andythenorth> and why wouldn’t that work? :P
07:31:20 <andythenorth> so nmlc or andythenorth is doing something wrong
07:31:56 <andythenorth> all will be revealed in next week’s episode
07:32:09 <andythenorth> along with a shocking twist you’d never have expected
07:32:22 * andythenorth -> important work to do
07:32:24 <andythenorth> things to delete
07:32:31 *** sla_ro|master has joined #openttd
07:36:53 <peter1138> Yup, that's a BAD FEATURE
07:37:16 <peter1138> As a generic string (lang bit bit 7 set) the feature is ignore. Except it's not totally ignored.
07:39:08 <andythenorth> do I need to care about any of the ‘totally’?
07:42:14 <peter1138> Hmm, OzTrans hasn't said I'm wrong yet :S
07:42:30 <andythenorth> I didn’t understand what he could say
07:42:37 <andythenorth> it’s just a call back, it needs a valid return
07:42:47 <andythenorth> what kind of handbags could we have about that?
07:47:05 *** FLHerne has joined #openttd
08:17:20 *** LadyHawk has joined #openttd
08:17:54 *** LadyHawk is now known as Guest3780
08:21:11 <andythenorth> there, deleted some stuff in one of our apps
08:42:58 *** andythenorth has left #openttd
08:43:13 *** trendynick has joined #openttd
09:07:31 <NGC3982> Is anyone here British? I'm trying to find a British person by phone number, but I can't seem to find any usable search engine or phone book.
09:08:18 <V453000> such a thing should thoretically not even exist, should it?
09:08:27 <V453000> I thought phone number is personal datra
09:11:29 <NGC3982> For us Swedes, we can search public web services for numbers, addresses and names. I guess we are the odd ones, maybe?
09:12:14 <planetmaker> pretty old news though :)
09:18:56 <FLHerne> NGC3982: You can search for phone number by name, but not vice versa. You might be able to find an electronic copy of the phone book for that area code?
09:25:49 <FLHerne> If it happens to be Huntingdonshire, I can look in my paper one :P
09:43:46 *** luaduck_zzz is now known as luaduck
09:44:23 *** luaduck is now known as luaduck_zzz
10:50:19 *** TheMask96 has joined #openttd
10:52:22 *** pthagnar has joined #openttd
11:04:54 *** UukGoblin has joined #openttd
11:10:54 *** [1]Tuhin has joined #openttd
11:15:53 *** [1]Tuhin is now known as Tuhin
11:16:54 *** Klanticus has joined #openttd
12:02:17 *** KWKdesign has joined #openttd
13:00:21 *** Midnightmyth has joined #openttd
13:10:30 *** oskari89 has joined #openttd
13:28:37 *** FLHerne has joined #openttd
13:45:21 <Alberth> what is param[133], and bit 59 in there?
13:46:30 <Alberth> It should be signals_on_traffic_side but no idea where that is in nfo
13:48:13 * Alberth has weird visions of exploding strawberries against the greenhouse glass :p
13:49:06 *** Midnightmyth has joined #openttd
13:49:52 <Alberth> pumpkins grown overnight :)
13:51:14 <V453000> without any green plants around them :D yeah
13:51:19 <V453000> got to fill those in later :d
13:51:23 <V453000> mainly changed the grass
13:51:28 <V453000> this one is a bit more interesting
13:51:36 <Eddi|zuHause> i thought those were tomatos, not strawberries
13:51:48 <V453000> they are, Alberth lives in skewed reality
13:53:36 <V453000> for once Alberth does not complain about sufficient amount of doodad objects btw? :D
13:54:04 <Alberth> maybe my skewed reality didn't see much empty space :p
14:18:26 <Alberth> yeah, I noticed, bit 59 of a 1 byte length variable :p
14:18:42 <Alberth> grfspecs seem a bit messed up
14:19:47 <Alberth> does nfo even have 64bit values?
14:21:41 <planetmaker> well, that variable is a 128bit size one :)
14:22:39 <Alberth> how do you get a bit from it?
14:24:45 <Alberth> I could try reading 88 or so, but that's another variable
14:34:56 *** KWKdesign has joined #openttd
14:44:52 <Eddi|zuHause> reading patch flags was a bit weird last time i looked
14:46:08 <Eddi|zuHause> i don't even remember why i looked
14:46:21 <Eddi|zuHause> but it was some horribly abusive construct that only reads single bits
14:48:12 <Alberth> I don't know what other ttdpatch flags are accessed by nml, but it looks totally broken
14:51:51 <Eddi|zuHause> as far as i understand it, you can only access it in action 6/7/9/D, not in varaction2
14:53:29 <Eddi|zuHause> you need to use condition type 00 or 01 (bit set/clear), and give the number of the bit as "value"
14:59:42 <Eddi|zuHause> Alberth: i think the only way to access it in a varaction2 is to store it in a parameter via action D, and then read that parameter in varaction2
15:00:13 <Eddi|zuHause> Alberth: or you just forbid reading TTDPatchFlags in varaction2 and throw an error.
15:01:29 <Alberth> it seems to do parameter reading or even expression value insertion with D/6 actions
15:01:55 <Alberth> so the former would be feasible, although I don't understand that part either
15:02:16 <Alberth> so far I am still stuck on getting bit 59 of an 8 bit variable :[
15:02:24 <Eddi|zuHause> can you even use action6 on a varaction2?
15:03:31 <Eddi|zuHause> Alberth: the "varsize" is not the size of the variable being read, but the size of the result being returned. does not have anything to do with each other
15:04:34 <kero> please, when updating some script on Bananas, should I supply it as .tar or .tar.gz ?
15:04:35 <Eddi|zuHause> so "test bit x of variable y" has the return type "byte"
15:04:54 <planetmaker> kero, should not matter
15:04:56 <Eddi|zuHause> kero: doesn't matter, bananas will compress it.
15:05:14 <Alberth> but how to express the bit?
15:05:42 <Alberth> there are no "get bit" operations, just the regular +, -, and shift stuff
15:05:43 <Eddi|zuHause> Alberth: the bit goes into the "value" parameter of the action 6/7/9/D
15:06:00 <Eddi|zuHause> Alberth: the condition type must be "is bit set" or "is bit clear"
15:06:38 <Alberth> 00 is assign and 01 is add
15:07:05 <Eddi|zuHause> hm, it's definitely in action7
15:08:00 <kero> Please, another question about Bananas. I don't remember how I should indicate dependencies. All I need, is clicking on them (one or more) to select them ?
15:09:01 <planetmaker> that only works, if you need the currently lates version of that dep on bananas
15:09:05 <planetmaker> otherwise you must use musa
15:09:51 *** sla_ro|master has joined #openttd
15:09:57 <Alberth> so you need 2 action7s? one for positive testing and one for negative testing, and jumping over the next sprites or not. If not jumping, do an action 6 to set a value
15:12:07 <Alberth> so how does that work if I toggle the value then?
15:13:03 <Alberth> 7 is only performed on loading afaik, yet I can toggle the signal side in-game
15:13:03 <Eddi|zuHause> toggling the value needs reloading the grfs
15:13:34 <Eddi|zuHause> this is all horribly wrong
15:14:23 <Eddi|zuHause> anyway, i think you can use two action D (one to set, one to clear) and one action 7 to skip the second one
15:18:42 <Eddi|zuHause> Action D1: set cache parameter
15:18:57 <Eddi|zuHause> Action 7: skip next action if ttdpatchflag is set
15:19:06 <Eddi|zuHause> Action D2: clear cache parameter
15:19:29 <Eddi|zuHause> from then on, you can read the cache parameter to check whether the flag is set
15:21:45 <Alberth> changing signal side doesn't sound like a safe action
15:22:38 <Alberth> yet I can change it without having developer rights
15:22:55 <Alberth> so I don't think it actually reloads newgrfs
15:26:15 <Alberth> your D/7/D sequence is a lot better though :)
15:34:23 <Eddi|zuHause> have i mentioned yet that the whole setting is wrong, anyway? it should a) be a company setting, and b) a direct left/right setting, not depending on traffic side
15:37:04 <kero> Mmh ... musa complains about "no license specified in the configuration file", though in the conf file there is "licence = GPL v2". What I am doing wrong ?
15:37:13 *** Hazzard has joined #openttd
15:37:59 <Eddi|zuHause> "license" vs. "licence"?
15:41:47 <Alberth> eddi: definitely one to nominate for kill -9 :p
15:43:04 <Alberth> btw do you have ideas how to make nml afloat again? I have been playing with c++ expressions, but it seems you'll end up rewriting nml completely
15:43:48 <Eddi|zuHause> i have not followed what you're doing
15:43:54 <Alberth> and it would be a long path before you have again something that runs and does something useful
15:44:30 <Alberth> mostly trying to understand how newgrf and nml works
15:45:08 <Alberth> and also find a way to move nml to something faster or more modular, or both
15:45:27 <Eddi|zuHause> the last thing i remember being discussed was expressions using some inefficient deep copy for no apparent reason
15:45:34 *** Klanticus_ has joined #openttd
15:45:55 <Alberth> but so far I failed to find a path that even looks reasonable doable
15:46:19 <kero> Still a question ... How do I indicate the target to musa ? giving the gz.tar file doesnt work. Giving the script directory neither. It complains about missing info.nut file :/
15:47:24 <Eddi|zuHause> Alberth: i'm not really fond of diving into that
15:47:59 <Alberth> fair enough, I just thought I'd ask, one never knows :) thanks
15:55:48 <juzza1> kero: musa only takes .grf and the like, not archive files
15:56:26 <kero> ok, but what do I do in case of a Game Script ?
15:56:32 *** Pensacola has joined #openttd
15:57:44 <kero> it's necesserily a directory or an archive :) Or something else I didn't thought about ?
15:59:00 <juzza1> can you put the files somewhere so i could test?
15:59:05 <juzza1> with -d of course, i wont upload anything
16:03:38 <kero> wait, the .ini file also.
16:06:14 <NGC3982> Sorry to say, i can't remember what the solution was to this the last time i asked
16:06:30 <NGC3982> But was there a way to simply syncronize -all- the online content at once?
16:07:28 <Alberth> not that I know, why would you want that?
16:07:42 <Alberth> most of the stuff is not for your style of playing
16:08:22 <NGC3982> If we change "-all-" to "lot's"? :-)
16:08:31 <NGC3982> It's just that i wish to skip the click fest, if possible.
16:08:37 <juzza1> kero: "python musa.py -c yourini.ini -d *.nut readme.txt" should work (run musa in the directory where you have all thos nut files)
16:08:53 <juzza1> and remove -d when you actually want to upload :)
16:09:12 <kero> thank you for the hint :)
16:09:23 *** Pensacola has joined #openttd
16:09:23 *** Klanticus_ has joined #openttd
16:09:23 *** Hazzard has joined #openttd
16:09:23 *** KWKdesign has joined #openttd
16:09:23 *** Eddi|zuHause has joined #openttd
16:09:23 *** tokai|mdlx has joined #openttd
16:09:23 *** Extrems has joined #openttd
16:09:23 *** bdavenport has joined #openttd
16:09:23 *** luaduck_zzz has joined #openttd
16:09:23 *** tyteen4a03 has joined #openttd
16:09:23 *** MTsPony has joined #openttd
16:09:23 *** theholyduck has joined #openttd
16:09:23 *** tycoondemon has joined #openttd
16:09:23 *** Xaroth|Work has joined #openttd
16:09:23 *** kais58__ has joined #openttd
16:09:23 *** Rubidium has joined #openttd
16:09:23 *** lobster has joined #openttd
16:09:23 *** blathijs has joined #openttd
16:09:23 *** tparker has joined #openttd
16:09:23 *** ccfreak2k has joined #openttd
16:09:23 *** davidstrauss_ has joined #openttd
16:09:23 *** eQualizer has joined #openttd
16:09:23 *** Belugas has joined #openttd
16:09:23 *** resistance.oftc.net sets mode: +ovov orudge orudge Belugas Belugas
16:09:23 *** Nothing4You has joined #openttd
16:09:23 *** jonty-comp has joined #openttd
16:09:25 <Alberth> NGC3982: you can do an upgrade of your newgrfs in the download window somewhere
16:10:05 <Alberth> kero: please document it somewhere appropriate, so people including you can find it the next time
16:12:58 * NGC3982 would gladly host an open sftp server for newgrf mirroring.
16:13:21 <MTsPony> is there a difference between videodriver dedicated and null?
16:14:11 <Alberth> probably, or one of them wouldn't exist :)
16:16:01 <kero> As a side note, still about musa. "musa.py --help" usage description indicates that there is an option "-q" (see [-dhqv]), though I don't see any in the detailed options.
16:16:02 <MTsPony> i guess, what is the difference? :p
16:16:20 <MTsPony> any difference in cpu cycles?
16:17:16 <Alberth> I think the difference is in the moment the drawing code gets skipped
16:17:31 <planetmaker> no. but it's a compile-time option. With dedicated you don't link to graphics driver libraries
16:17:57 <Alberth> oh :) thanks planetmaker
16:18:22 <planetmaker> I hope that's right. Sounds plausible though :P
16:19:28 <MTsPony> so assuming --enable-dedicated is used, putting null or Dedicated for videodriver makes no difference ?
16:21:06 <planetmaker> I don't know, but that will be my guess w/o looking at source :)
16:23:14 <planetmaker> so what about the outstanding edit / supply of the required patch file, MTsPony ?
16:23:26 <planetmaker> "sorry" without supplying it is not really enough
16:24:22 <juzza1> kero: also -r is not mentioned on the "Usage" line of --help, someone must've typoed r to q as -q is not in the source
16:24:56 <planetmaker> svn diff > patch.diff? upload, edit post done?
16:25:00 <planetmaker> what work is there to do?
16:25:31 <planetmaker> if it's more work, you do something wrong
16:27:08 <planetmaker> I hope you use svn (or hg or git) to work on the code, you do, yes?
16:27:11 <kero> ah, yes, that's the think
16:27:46 <MTsPony> perhaps you should relax a bit , i said i was going to solve it, i took the link down
16:28:36 <planetmaker> that's not my point :) I try to understand why it's so hard and would like try to help you
16:29:50 <MTsPony> so hard? some people have lives beside a computer i had more pressing matters which required my attention
16:30:23 <planetmaker> compiling the thing is way harder than getting a diff. So it was 2 minutes of work missing when you actually created them
16:30:40 <planetmaker> "life besides computer" is an extremely bad excuse for not caring about a license
16:31:36 <planetmaker> but if you want to get me actually annoyed, that's fine, too.
16:31:50 <MTsPony> You're impossible. lol
16:32:50 <planetmaker> right. I'll make a note in not trying to be understanding but writing DMCA right away?
16:33:29 <MTsPony> I told you i wouod be supplying the diff file, until then i took the links down
16:33:36 <MTsPony> isnt that,good,enough?
16:34:23 <MTsPony> and i also apologized and i would correct the issue.
16:35:56 <planetmaker> and I just tried to understand how this keeps happening :) Not you but generally. Wondering what information or awareness was missing from the start. So that I can learn from that. And maybe better educate. Or improve documentation. Or whatever
16:36:00 *** Jinassi has joined #openttd
16:36:43 <MTsPony> most people are just players that try to have fun and do normally not think about licensing
16:37:05 <MTsPony> and if you want to know the delay i had to go to the vet
16:39:13 <Jinassi> should i edit the wiki page or is "coming soon" not self-explanaory enough?
16:39:26 <planetmaker> nah, don't bother
16:39:31 <Jinassi> pardon my typos, i have an urge to hit things now
16:47:44 *** Klanticus has joined #openttd
16:50:07 *** LadyHawk has joined #openttd
16:57:36 *** Progman has joined #openttd
17:02:25 *** Klanticus has joined #openttd
17:29:30 *** andythenorth has joined #openttd
17:46:03 <DorpsGek> Commit by translators :: r26703 /trunk/src/lang (3 files) (2014-07-25 17:45:54 UTC)
17:46:04 <DorpsGek> -Update from WebTranslator v3.0:
17:46:05 <DorpsGek> belarusian - 69 changes by KorneySan
17:46:06 <DorpsGek> norwegian_bokmal - 1 changes by
17:46:07 <DorpsGek> russian - 3 changes by KorneySan
17:46:08 <DorpsGek> slovenian - 26 changes by matej1245
18:13:09 *** andythenorth has joined #openttd
18:16:46 <DorpsGek> Rubidium: DanMack was last seen in #openttd 2 days, 0 hours, 23 minutes, and 50 seconds ago: <DanMacK> @seen andythenorth
18:16:54 <Rubidium> there... now you don't need to do it ;)
18:22:46 <andythenorth> @seen is a BAD FEATURE
18:22:46 <DorpsGek> andythenorth: seen [<channel>] <nick>
18:23:36 <andythenorth> spose I’d better fix my fucked up grf :|
18:26:43 *** frosch123 has joined #openttd
18:29:31 <andythenorth> type_name: string(STR_CARGO_NAME_ALCOHOL);
18:29:39 <andythenorth> (has a property block around it)
18:29:45 <andythenorth> puzzlingly results in silly nfo
18:34:12 <andythenorth> nmlc provides a string ID correctly in the dummy cb afaict
18:34:14 <andythenorth> \wx806E \dx0000006E \dx0000006E // 110 .. 110: return string(STR_CARGO_NAME_ALCOHOL);
18:34:29 <andythenorth> but it doesn’t seem to write a corresponding action 4 to provide the strings
18:34:34 <andythenorth> it does do this for trains
18:35:08 <andythenorth> and it’s resolving type_name: string(STR_CARGO_NAME_ALCOHOL); to \wxDC00
18:35:33 <andythenorth> I bet it’s my code, but I’m wondering if I have to trawl nmlc
18:37:15 <andythenorth> grfcodec can’t build that to a valid grf, missing header info
18:37:21 <andythenorth> but that’s not the issue
18:42:47 <frosch123> Alberth: the ttdpflags are the reason why you cannot change signal side in multiplayer
18:43:04 <frosch123> in singleplayer changing the side only takes effect when reloading the game
18:43:12 <frosch123> in multiplayer it is forbidden to prevent desyncs
18:43:21 <frosch123> anyway, ttdpflags are no variable
18:43:42 <frosch123> they are an abomination
18:44:13 <Alberth> better drop them perhaps?
18:44:29 <Alberth> or at least, in nml :)
18:44:46 <frosch123> we can replace them with other variables
18:44:52 <frosch123> which contain only useful bits
18:45:03 <frosch123> oh. btw. majority of ttdp flags are undefined
18:45:13 <Alberth> I'd settle for a sane way to access it already :p
18:45:18 <frosch123> ttdp has a nice table, where you can assign fixed bits to certain config options
18:45:30 <frosch123> the other config options are distributed randomly on the other bits
18:45:55 <frosch123> which means that in ttdp you can only test a ttdpflag after testing the ttdp version aganist the version when the bit got a fixed position
18:46:35 <frosch123> anyway, most ttdp flags are pretty useless
18:47:04 <frosch123> however the signal side flag is uses to choose signal variants by some grf, because they are too lazy to add a parameter
18:47:25 <frosch123> or because it was in base graphics which do not have parameters
18:47:31 <frosch123> but well, all bollocks :)
18:47:47 <Alberth> not a good reason for using the flag indeed
18:55:24 <andythenorth> so nmlc is not writing out any DCxx texts
18:55:33 <andythenorth> but it is assigning cargo props to use DCxx texts
18:55:44 <andythenorth> (all DC00, whic looks wrong, but eh)
18:58:06 *** montalvo has joined #openttd
18:59:17 *** Myhorta has joined #openttd
19:01:21 <andythenorth> if the cargo is in nml containing a header block it works
19:01:43 <andythenorth> not sure if that’s significant or not
19:02:10 <planetmaker> string processing might be only called when actually writing the grf
19:02:14 <andythenorth> maybe it’s a co-incidence that trains work
19:02:17 <planetmaker> (as a wild guess)
19:02:25 <andythenorth> trains write an action 4, cargos use action 0 props for strings
19:03:57 <planetmaker> hm, possibly cargoes can have a fixed amount of strings. Their IDs are hard-coded. While there's an undetermined amount of trains. Thus action4 with variableID needed
19:06:34 <andythenorth> this is tractable
19:06:53 <andythenorth> not sure how I fix the strings that need substrings though
19:06:55 <andythenorth> one step at a time
19:09:13 *** shirish_ has joined #openttd
19:12:04 *** montalvo has joined #openttd
19:31:26 <peter1138> Hmm, sprite sorting issues.
19:34:09 <peter1138> V453000, will you make animations only happen when industries are producing?
19:53:12 <V453000> peter: yeah once I manage to code it :)
20:09:28 <planetmaker> V453000, is the intention to run it once upon production?
20:10:12 <frosch123> none of the animations have a "once"
20:10:41 <frosch123> they all looked continous to me
20:12:00 <planetmaker> in yeti? Currently surely yes :)
20:12:23 <frosch123> maybe he wants to control the speed depending on the production rate :p
20:12:41 <planetmaker> I even once suggested that :P
20:12:43 <frosch123> may require new "blur" sprites
20:12:48 <andythenorth> maybe he should finish YETI :P
20:12:52 <andythenorth> so I can stop using FIRS
20:13:06 <andythenorth> we’ll need YETI-GS
20:13:19 <frosch123> andythenorth: i wonder how we get v to make a ship set though :p
20:13:40 <planetmaker> we also need a RV set
20:14:15 <planetmaker> besides HEQS - which is not a full set - I'm always left a bit helpless as to which to choose
20:16:03 <Eddi|zuHause> <frosch123> andythenorth: i wonder how we get v to make a ship set though :p <-- only if you make him forget he ever coded wetrails
20:16:19 <andythenorth> planetmaker: OGFX+ vehicles
20:16:52 <planetmaker> well, yes. But else?
20:17:57 <planetmaker> uhm... it's not always satisfyable.
20:18:11 <planetmaker> And I hate articulated vehicles not being able to overtake :P
20:19:15 <andythenorth> Road Hog is mostly articulated
20:19:38 <Eddi|zuHause> planetmaker: implement better road traffic ;)
20:19:40 <planetmaker> it's not something I can hold against a decent RV grf
20:19:41 <frosch123> alternatively we could make rv crash slower vehicles in front
20:19:55 <frosch123> would also nerve drive-through stops
20:20:44 <andythenorth> so no nut-like RVs
20:22:42 <Eddi|zuHause> i think road vehicles should behave like this: RV reserves <space> ahead, depending on speed. if reservation fails, road vehicle enters the "follow" mode, where it adjust its speed and distance to the vehicle ahead (reference stored in the vehicle). if RV is in "follow" mode and <speed difference> > x, enter "overtake" mode, and try to reserve <distance> depending on speed difference. if reservation fails, return to follow mode
20:23:54 <Eddi|zuHause> need bits in the map to store the reservation
20:24:25 <frosch123> to make other rv stop or what?
20:25:40 <Eddi|zuHause> reservation's main purpose in the overtake mode is to detect vehicles coming in the opposite direction
20:26:29 <Eddi|zuHause> so road in the opposite lane must be free for at least <distance> before attempting overtaking
20:27:08 <frosch123> and why does that need map bits?
20:27:16 <frosch123> because that's just what the current overtake code does
20:27:47 <Eddi|zuHause> frosch123: but the current overtake code's <distance> is too low
20:28:04 <Eddi|zuHause> frosch123: and it has this weird "no more than two vehicles on one tile" rule
20:28:25 <Eddi|zuHause> which does nothing useful
20:29:35 <Ailure> could be a little more liberal on one-way roads too
20:30:19 <Eddi|zuHause> frosch123: if all road vehicles which are not in "follow" mode reserve <n> ticks ahead, any vehicle can make overtaking attempts that are guaranteed to finish in <n-1> ticks
20:31:31 <Eddi|zuHause> and you can calculate this time from current vehicle's speed and ahead vehicle's speed, plus some fuzz for acceleration and slopes
20:32:52 <frosch123> reservation in map array won't work
20:33:15 <Eddi|zuHause> it works for trains
20:33:33 <frosch123> a tile is reserved by multiple rv, and if you try to track which rv still reserve it, you do not need it in the first palce
20:33:51 <Eddi|zuHause> except you alter the behaviour when you encounter a reservation
20:34:21 <Eddi|zuHause> once you encounter a reservation, you store that reference vehicle in the current vehicle
20:34:46 <Eddi|zuHause> creating a linked list of "following" vehicles
20:38:27 <andythenorth> planetmaker: Road Hog has ‘interesting’ graphics
20:38:36 <Eddi|zuHause> frosch123: maybe reservation needs to be more fine-grained, i.e. on step level (16 per lane and tile)
20:38:41 <andythenorth> otherwise known as ‘unfinished’
20:38:56 <Eddi|zuHause> but that's potentially 64 bit per tile
20:39:28 <Eddi|zuHause> on an X crossing there are 4 lanes
20:39:55 <frosch123> you even want to do it over crossings? :p
20:40:15 <Eddi|zuHause> yes, otherwise "onramps" for your highway break overtaking
20:40:16 <frosch123> overtaking on the inside of a curve :)
20:40:21 <Eddi|zuHause> which they shouldn't
20:40:52 *** montalvo has joined #openttd
20:41:18 <Eddi|zuHause> this has the side effect that you can block a crossing for cross vehicles. allowing to make "traffic lights" useful
20:43:46 <Eddi|zuHause> frosch123: alternatively, you need one bit to reserve a lane, and one bit to define the direction it is reserved, making 8 bit per tile. but then you have to check on unreserving a lane whether there is a following vehicle that needs this lane
20:44:04 <Eddi|zuHause> which could be achieved by making a doubly-linked-list
20:44:04 <MTsPony> Error: Assertion failed at line 1524 of /home/marctraider/trunkreddit4/trunk/src/saveload/saveload.cpp: IsVariableSizeRight(sld)
20:44:19 <MTsPony> anyone recognize this error?
20:44:40 <Eddi|zuHause> MTsPony: your patch is broken
20:44:41 <frosch123> try r26700 or newer
20:45:20 <MTsPony> since then it started happening :(
20:46:11 <frosch123> well, it was worse before :p
20:46:21 <frosch123> it just did not detect it
20:47:42 <MTsPony> you think 26700+ fixes it?
20:48:07 <frosch123> it fixes one of those issues
20:48:10 <frosch123> but there can be more
20:48:25 <frosch123> the bugs are not new, but the detection is
20:48:25 <planetmaker> so... are yeti dudes like cargo or passengers? :P
20:48:27 <MTsPony> ah. never had it with 26335, thats a while back tho
20:48:37 <frosch123> planetmaker: like supplies
20:48:53 <planetmaker> I'm considering the choice of engine from nuts currently :)
20:49:11 <frosch123> oh, they take the flat bed van iirc
20:51:58 <planetmaker> cool. yetis look like rubber ducks
20:53:54 *** zeknurn has joined #openttd
21:01:32 <planetmaker> good night for today
21:09:33 <MTsPony> Mhh so there have recently been working on savegame.cpp as when i downgrade from latest revision to 26694, the one we use to apply the patch against, it changes that file i.e.
21:09:41 <MTsPony> is that error we get related to these changes?
21:10:04 <MTsPony> or did the patch we use became 'incompatible' somehow?
21:11:00 <Eddi|zuHause> MTsPony: the change was to detect mistakes in the saveload code more fiercely
21:11:24 <Eddi|zuHause> MTsPony: so if you hit this assert, it was broken before, but you wouldn't notice it (or only when it desyncs later)
21:11:49 <MTsPony> ah. but we had 0 desyncs ever
21:11:57 <MTsPony> with 26335 (same patch give or take)
21:11:59 <Eddi|zuHause> that doesn't mean anything
21:12:22 <Eddi|zuHause> MTsPony: like frosch123 said, even trunk hat instances of hitting this assert
21:12:41 <MTsPony> in what revision did this more fierce detection got introduced?
21:20:55 <MTsPony> best i could find was 2 months ago fonsinchen (svn r26551) -Fix: document revision number for latest savegame version change
21:22:23 <MTsPony> frosch mentioned updating to 26700>
21:22:28 <MTsPony> does 4 days ago rubidium (svn r26700) -Fix [FS#6066]: incorrect saving of order backups
21:22:31 <MTsPony> have anything to do with it?
21:26:15 <Eddi|zuHause> wll have you tried?
21:27:12 <MTsPony> that must be the one that does the extra checking
21:32:02 <Rubidium> not getting a desync doesn't mean the code is desync safe
21:34:44 <Rubidium> most of the cases that trigger the assertion will not trigger a desync for most people. Mostly because the values that are stored are often close to 0, so saving two bytes too little is generally not causing a problem, but it could. The most troublesome these savegame issues are for big endian systems as for them the important data gets lost (the numbers near to 0)
21:37:34 <MTsPony> well since i dont really know how i can manually trigger this same thing, or perhaps manual saving could trigger it id have to check
21:39:19 <Rubidium> was it a server or network game client that gave the error, or was a game previously saved in MP loaded?
21:41:00 <MTsPony> we have both the client and sever yeah? same revision same patch. I made a heightmap in scenario editor, then made a savegame out of it, dumped it to linux server like i always did, then load it up
21:41:10 <MTsPony> its the server that keeps doing it
21:41:46 <MTsPony> im not sure, but could 'aggresive' cxxflags be the cause of this?
21:42:11 <Rubidium> well, r26700 fixes an issue that only triggers in the cases I was just asking about... so it could already been fixed
21:42:24 <Rubidium> alternatively it could be the patch you're using
21:42:46 <Eddi|zuHause> Rubidium: i vaguely remember that the saveload code actually asserted on writing too large values.
21:43:03 <MTsPony> no its from a heightmap, and i made the savegame in single player mode before i copied it to server
21:43:14 <MTsPony> so i guess 26700 wouldnt apply to fix it :(
21:43:31 <Rubidium> so the server crashes immediately when loading the savegame?
21:43:37 <MTsPony> nope. just after a while.
21:43:52 <MTsPony> i suspect when it autosaves, but it doesnt seem to do it with every autosave
21:43:54 <Rubidium> so, it is a server that is crashing (during saving)
21:44:10 <Rubidium> and... a joining client implies saving
21:45:09 <Eddi|zuHause> in ancient times, some versions of daylength caused asserts on saving, because the the AI start counter would overflow
21:46:27 <Rubidium> that's a, generally, different kind of issue
21:47:10 <Rubidium> the assert that I added recently is about reading fewer bytes from the in-memory variable than the size of the actual variable
21:47:47 <Rubidium> as a result, in case of big endian, the "wrong" bit of memory will be read/written
21:49:10 <Rubidium> but before you figure that out in a desync hunt you'd be days further, so it was a good idea to add a check that just triggers whenever it can (too bad it couldn't be done statically)
21:50:00 <MTsPony> Perhaps aggressive CXXFLAGS could cause it?
21:50:27 <Rubidium> that's unlikely, but I won't rule it out
21:50:31 <Rubidium> anyhow... I agree with the sentiment of frosch and Wolf
22:10:47 <peter1138> Hmm, my ipsec connection appears to drop every so often :(
22:10:56 <peter1138> Keep-alives are enabled :S
22:11:19 <peter1138> Oh, maybe it's just the SA lifetime.
22:11:55 <peter1138> Typical, thinking of reasons just after telling IRC :p
22:33:25 *** glx is now known as Guest3841
22:48:47 <Eddi|zuHause> that's an actual psychological effect. saying things outloud activates some connections in your brain
23:24:30 *** KWKdesign has joined #openttd
23:33:25 *** luaduck_zzz is now known as luaduck
23:41:17 *** luaduck is now known as luaduck_zzz
continue to next day ⏵
