Log for #openttd on 24th September 2018:
Times are UTC Toggle Colours
00:01:13  *** greeter has quit IRC
00:12:34  *** Gustavo6056 has joined #openttd
00:18:13  *** Gustavo6046 has quit IRC
00:18:13  *** Gustavo6056 is now known as Gustavo6046
00:25:12  *** Flygon has joined #openttd
00:28:39  *** ccfreak2k has quit IRC
00:30:51  *** ccfreak2k has joined #openttd
00:41:12  *** Thedarkb-T60 has quit IRC
01:12:25  *** Gustavo6056 has joined #openttd
01:18:48  *** Gustavo6046 has quit IRC
01:18:48  *** Gustavo6056 is now known as Gustavo6046
01:36:33  *** cHawk has quit IRC
01:48:31  *** acklen has quit IRC
01:48:36  *** acklen has joined #openttd
01:55:28  *** glx has quit IRC
02:06:44  *** cHawk has joined #openttd
02:09:47  *** mindlesstux has quit IRC
02:09:58  *** mindlesstux has joined #openttd
02:12:43  *** Gustavo6056 has joined #openttd
02:19:53  *** Gustavo6046 has quit IRC
02:19:53  *** Gustavo6056 is now known as Gustavo6046
02:59:15  *** Samu has quit IRC
03:12:50  *** Gustavo6056 has joined #openttd
03:13:42  *** haudrauf has quit IRC
03:15:06  *** lethosor has quit IRC
03:15:17  *** lethosor has joined #openttd
03:16:30  *** haudrauf has joined #openttd
03:19:28  *** Gustavo6046 has quit IRC
03:19:28  *** Gustavo6056 is now known as Gustavo6046
03:30:27  *** milek7_ has quit IRC
03:30:28  *** milek7 has joined #openttd
04:13:00  *** sla_ro|master has joined #openttd
04:13:02  *** Gustavo6056 has joined #openttd
04:14:56  *** ToBeFree has joined #openttd
04:15:12  *** ToBeFree is now known as ToBeCloud
04:19:02  *** Gustavo6046 has quit IRC
04:19:02  *** Gustavo6056 is now known as Gustavo6046
04:22:07  *** rocky113844 has quit IRC
04:24:12  *** sla_ro|master has quit IRC
04:57:02  *** snail_UES_ has quit IRC
05:06:27  *** Westie has quit IRC
05:07:29  *** Westie has joined #openttd
05:13:01  *** Gustavo6056 has joined #openttd
05:17:38  *** Gustavo6046 has quit IRC
05:17:38  *** Gustavo6056 is now known as Gustavo6046
05:40:51  *** masse has quit IRC
05:41:13  *** masse has joined #openttd
05:44:26  *** Antheus has quit IRC
05:44:33  *** Antheus has joined #openttd
06:01:39  *** DorpsGek_II has quit IRC
06:13:09  *** Gustavo6056 has joined #openttd
06:14:00  *** andythenorth has joined #openttd
06:19:53  *** Gustavo6046 has quit IRC
06:19:53  *** Gustavo6056 is now known as Gustavo6046
06:39:15  *** Alkel_U3 has quit IRC
06:39:28  *** Alkel_U3 has joined #openttd
06:42:33  *** andythenorth has quit IRC
06:50:19  *** andythenorth has joined #openttd
07:07:53  *** Mek has quit IRC
07:09:52  *** andythenorth has quit IRC
07:12:03  *** IgnoredAmbience has quit IRC
07:12:17  *** IgnoredAmbience has joined #openttd
07:13:17  *** Gustavo6056 has joined #openttd
07:17:17  *** hph^ has quit IRC
07:17:35  *** hph^ has joined #openttd
07:18:22  *** Gustavo6046 has quit IRC
07:18:22  *** Gustavo6056 is now known as Gustavo6046
07:19:13  *** Mek has joined #openttd
07:37:09  *** andythenorth has joined #openttd
08:13:30  *** Gustavo6056 has joined #openttd
08:18:20  *** andythenorth has quit IRC
08:18:22  *** Gustavo6046 has quit IRC
08:18:22  *** Gustavo6056 is now known as Gustavo6046
08:28:59  *** wodencafe has quit IRC
08:29:22  *** wodencafe has joined #openttd
09:13:37  *** Gustavo6056 has joined #openttd
09:20:32  *** Gustavo6046 has quit IRC
09:20:32  *** Gustavo6056 is now known as Gustavo6046
09:55:54  *** m3henry has joined #openttd
10:13:39  *** Gustavo6056 has joined #openttd
10:20:03  *** Gustavo6046 has quit IRC
10:20:03  *** Gustavo6056 is now known as Gustavo6046
10:53:51  *** Samu has joined #openttd
10:54:28  <Samu> hi
11:01:17  *** greeter has joined #openttd
11:03:20  *** masse is now known as MasseR
11:13:57  *** Gustavo6056 has joined #openttd
11:20:07  *** Gustavo6046 has quit IRC
11:20:07  *** Gustavo6056 is now known as Gustavo6046
11:47:47  <Samu> i'm inventing a new cost!
11:47:51  <Samu> drive through
11:48:15  <Samu> _cost_drive_through = null;    ///< The extra cost if a road tile is part of a drive through road station.
12:13:59  *** Gustavo6056 has joined #openttd
12:20:42  *** Gustavo6046 has quit IRC
12:20:42  *** Gustavo6056 is now known as Gustavo6046
12:52:13  *** snail_UES_ has joined #openttd
13:01:08  *** snail_UES_ has quit IRC
13:14:01  *** Gustavo6056 has joined #openttd
13:19:52  *** Gustavo6046 has quit IRC
13:19:52  *** Gustavo6056 is now known as Gustavo6046
13:41:49  *** nielsm has joined #openttd
13:49:34  *** Gja has joined #openttd
13:53:54  <Samu> is this explanation confusing:
13:53:55  <Samu> 	_cost_no_existing_road = null; ///< The cost that is added to _cost_tile if no road connection exists between two tiles. Added twice when the tile to enter has no road, no bridge and no tunnel.
13:54:37  <Samu> how to english?
13:56:08  <m3henry> Perhaps "Cost is doubled when..."
13:56:18  <Samu> ok
13:56:51  <Samu> done
13:57:05  <m3henry> Also is 'null' appropriate? Doesn't sound like a pointer to me
13:57:26  <Samu> yes, at this place
13:57:48  <Samu> the constructor comes next
13:57:51  <Samu> with this._cost_no_existing_road = 40;
13:59:29  *** rocky113844 has joined #openttd
13:59:46  <Samu> but the explanation per se, does it sound confusing?
14:00:31  <m3henry> Is it no_bridge && no_road && no_tunnel?
14:00:48  <Samu> sec,
14:01:18  <Samu> it's this
14:01:36  <Samu> guess it's a ||
14:03:05  <Samu> IsRoadTile includes drivethrough already
14:03:58  <Samu> and since road vehicles can walk on competitor's drive throughs, it's not needed to check if the drive through is mine
14:05:25  <m3henry> I would say that the if statement is a bit unweildy
14:10:31  <Samu> let me try something different
14:12:03  <Samu>
14:12:05  <Samu> is this better?
14:14:00  <Samu> maybe not
14:14:25  *** Gustavo6056 has joined #openttd
14:14:42  <Samu>
14:14:50  <Samu> what would you do m3henry
14:15:04  <m3henry> This may be more readable:
14:16:17  *** Flygon has quit IRC
14:19:26  <Samu> which one would be computed with less operations?
14:19:27  *** Gustavo6046 has quit IRC
14:19:27  *** Gustavo6056 is now known as Gustavo6046
14:19:30  <m3henry> By breaking the statement up, you can explain the logic in the variable names
14:19:57  <Samu> this function is called a bajillion times for almost every tile while pathfinding
14:20:09  <Samu> so, it's also important to get the same result in less ops
14:20:25  <m3henry> Have you benchmarked it?
14:20:37  <Samu> not really, but
14:20:43  <m3henry> but benchmark it
14:20:51  <Samu> i saw it in action, it's really intensive
14:21:01  <Samu> debug signs everywhere
14:21:28  <Samu> not sure how i'm gonna benchmark such thing
14:22:16  <Samu> i notice, for example, a high value on here will make the pathfinder check more tiles
14:23:24  <m3henry> An aggressive optimizing compiler is likely to emit similar if not identical machine code for those statements
14:24:07  <m3henry> It's not something I would worry about
14:24:32  <Samu> there's a max #opcodes :o
14:24:41  <Samu> but oki
14:24:55  <m3henry> Have you got the most efficent algorithm?
14:26:16  <Samu> no idea, i was just trying to make the costs look consistent
14:27:22  <Samu> i suppose the original code was faster, for a reason alone, it only tests bridges on sloped tiles
14:27:38  <Samu> my changes make it test bridges on every kind of slope
14:27:47  <Samu> flat slopes
14:28:06  <m3henry> If it is more correct, then you'll have to take that hit
14:28:13  <Samu> pretty much yes
14:28:41  <m3henry> perhaps you can amorrtize the cost over n cycles
14:30:18  <Samu> the original costs was doing DistanceManhattan multiple times for costs
14:30:23  <Samu> erm...
14:30:34  <Samu> repeating the same
14:30:47  <Samu> so i turned it into a local dist = DistanceManhattan
14:30:58  <Samu> i suppose it would fasten up
14:31:24  <Samu> not really sure how the opcodes measure
14:31:46  <Samu> but i expect there would be less operations needed
14:33:12  <Samu> here's the original
14:39:44  *** TinoDidriksen has quit IRC
14:39:56  *** TinoDidriksen has joined #openttd
14:40:35  *** TinoDidriksen is now known as Guest1243
14:41:43  *** Guest1243 is now known as TinoDidriksen
14:43:12  <Samu> sec
14:43:55  <Samu>
14:43:58  <Samu> maybe i can measure :p
14:44:04  <Samu> let me experiment this
14:44:25  *** Alberth has joined #openttd
14:44:25  *** ChanServ sets mode: +o Alberth
14:45:47  <Alberth> o/
14:47:51  <Samu> hi
14:49:38  <Samu> how do i actually measure
14:49:53  <Samu> without influencing the measure myself with debug messages
14:50:08  <Samu> seems that's something not possible to do
14:50:37  *** m3henry has quit IRC
14:53:11  <Samu>
14:53:12  <Samu> lel
14:53:59  <Samu> why do i get a big negative value at times?
14:54:06  <Samu> max opcodes atm is 10000
14:56:56  <Samu> gonna try ur code
14:59:05  <Samu> 49
14:59:10  <Samu> my code wins
14:59:12  <Samu> with 48
14:59:21  <Samu> lol
14:59:54  <Samu> oh, he left
15:08:33  <Samu> got a 47 with my original code
15:10:26  <Samu> if there is a connection: 12
15:11:31  <Samu> if there is a road on new_tile: 26
15:11:33  *** sla_ro|master has joined #openttd
15:11:45  <Samu> if there is no road on new_tile: 47
15:12:04  <Samu> no road on new_tile is the most common occurence
15:14:43  *** Gustavo6056 has joined #openttd
15:21:27  *** Gustavo6046 has quit IRC
15:21:27  *** Gustavo6056 is now known as Gustavo6046
15:26:07  *** WWacko1976-work has quit IRC
15:34:11  *** Wormnest has joined #openttd
15:44:38  *** Gja has quit IRC
15:46:36  *** quiznilo has left #openttd
15:46:54  <Samu> 12 / 26 / 47 vs 12 / 23 / 48 vs 12 / 37 / 49
15:49:22  *** Thedarkb-T60 has joined #openttd
15:56:12  <Samu> Alberth: (12/26/47) vs (12/23/48) vs (12/37/49)
15:56:23  <Samu> number of opcodes
15:57:53  <Samu> x/y/z : x opcodes if there is a connection, y opcodes if there is a road on new_tile, z opcodes if there is no road on new_tile
15:58:36  <Samu> i'm undecided between the first and second codes
15:58:43  <Samu> what is preferible
15:58:55  <Alberth> throw a die
15:59:29  <Samu> I'm more inclined towards 12/26/47
16:00:13  <Alberth> you get opcodes in the 10,000s right?
16:00:22  *** Gja has joined #openttd
16:00:24  <Samu> yes, the default is 10000
16:00:26  <Alberth> not sure how significant 1 op is
16:01:26  <Samu> it's calculated for every tile neighbour stuff
16:02:57  <Samu> @calc 10000 / 47
16:02:57  <DorpsGek> Samu: 212.765957447
16:03:00  <Samu> @calc 10000 / 48
16:03:00  <DorpsGek> Samu: 208.333333333
16:03:01  <Samu> @calc 10000 / 49
16:03:01  <DorpsGek> Samu: 204.081632653
16:04:00  <Alberth> you're avoiding doing double calculations?
16:05:26  <Alberth> that 212 vs 208 isn't true, you also do other things than just computing a tile status
16:07:06  <Samu> z is the most common occurence t.t
16:07:32  <Samu> building on a empty tile
16:08:13  <Samu> y happens when it connects to an existing road
16:08:31  <Samu> x happens when it's already connected
16:10:03  <Samu> y is probably the most rare
16:10:33  <Samu> so i'm inclined towards that z =47
16:10:52  <Samu> there's just too many empty tiles on the map
16:13:45  <Samu> haven't tested computing stations
16:13:51  <Samu> would probably go over 50
16:14:02  <Samu> stations or depot
16:14:29  *** Gustavo6056 has joined #openttd
16:14:40  <Samu> there's much less stations/depots on the map than there is roads
16:15:38  <Samu> unless you can come with a more efficient way to do these checks
16:16:30  <Samu> I'll settle for for now
16:21:58  *** Gustavo6046 has quit IRC
16:21:58  *** Gustavo6056 is now known as Gustavo6046
16:24:32  *** Gustavo6056 has joined #openttd
16:25:39  *** HerzogDeXtEr has joined #openttd
16:30:09  *** Thedarkb-T60 has quit IRC
16:31:18  *** frosch123 has joined #openttd
16:31:28  *** Gustavo6046 has quit IRC
16:31:28  *** Gustavo6056 is now known as Gustavo6046
16:57:17  <Samu>
16:57:19  <Samu> heh
16:58:09  <Samu> just 1 tile
16:58:26  <Samu> 1 neighbour
17:01:30  <Samu> hmm _cost_slope is a bit expensive
17:02:55  *** lethosor_ has joined #openttd
17:05:56  *** rocky113844 has quit IRC
17:05:57  *** lethosor has quit IRC
17:05:57  *** cHawk has quit IRC
17:05:57  *** ccfreak2k has quit IRC
17:06:30  *** rocky113844 has joined #openttd
17:06:35  *** ccfreak2k has joined #openttd
17:06:35  *** cHawk has joined #openttd
17:14:42  *** Gustavo6056 has joined #openttd
17:21:39  *** andythenorth has joined #openttd
17:22:02  *** Gustavo6046 has quit IRC
17:22:02  *** Gustavo6056 is now known as Gustavo6046
17:23:03  <andythenorth> moin
17:24:16  *** Progman has joined #openttd
17:29:56  *** gelignite has joined #openttd
17:36:36  <Samu> the slope checking is expensive t.t
17:36:44  <Samu> grr, how to fix
17:36:54  <Samu> must dissect it
17:43:43  <andythenorth> is it stuff?
17:44:02  <andythenorth> I have played tanks enough
17:44:07  <andythenorth> are we making the game?
17:48:14  <Samu>
17:48:31  <Samu> this part cost 80-107 ops
17:48:34  <Alberth> moin
17:48:43  <Samu> how to make it more efficient?
17:48:44  <Samu> halp
17:48:49  <Alberth> at least samu is making the game :)
17:49:18  <Samu> and if combined with the isbridge and istunnel checks it can go to 140s
17:52:14  *** Alberth has left #openttd
18:00:35  <nielsm> Samu try replacing the first 8 lines with this, it might be less ops:
18:05:11  <Samu> woah, let me try that
18:05:32  <LordAro> looks nicer too :p
18:08:41  <Samu> i reduced from 107 to 98
18:08:46  <Samu> now testing nielsm code
18:09:23  <Samu> 86
18:09:26  <Samu> woah, thx
18:10:13  <Samu> i hope it's doing the same as before, i didn't really verify that
18:10:21  <nielsm> it should be
18:10:48  <nielsm> basically eliminates some redundant calls to the library (they might be expensive, might not) and some shared expression elimination
18:14:44  *** Gustavo6056 has joined #openttd
18:17:15  <Samu> tried this
18:17:16  <Samu> local NW = middle - AIMap.GetMapSizeX();
18:17:25  <Samu> local SE = middle + AIMap.GetMapSizeX();
18:17:29  <Samu> got 89
18:17:30  <andythenorth> so nielsm
18:17:34  <Samu> so yeah, it's worse
18:17:37  <andythenorth> what silly thing shall we do next? o_O
18:18:07  <LordAro> the same thing we do every night
18:18:57  <andythenorth> look at PRs and wonder how to get them approved? o_O
18:20:24  <LordAro> yup!
18:20:32  <nielsm> I was working on those GS controls of industry production levels
18:20:51  <nielsm> got slightly sidetracked by a swathe of cargo types :P
18:21:04  <LordAro> frosch123: you should merge some PRs
18:21:38  <andythenorth> I wondered about starting a patchpack
18:21:41  <LordAro> andythenorth: ^ that's how it works, right?
18:21:58  *** Gustavo6046 has quit IRC
18:21:58  *** Gustavo6056 is now known as Gustavo6046
18:22:14  <andythenorth> LordAro: sometimes :P
18:22:24  <andythenorth> hmm
18:22:33  <andythenorth> I think my C++ is too piss poor for me to maintain a PP
18:22:44  <andythenorth> and I don't think it will improve fast enough to be worth it
18:23:20  <LordAro> for the most part there's very little involved
18:23:24  <LordAro> only when they overlap
18:23:34  <LordAro> most of it is git skills >:)
18:23:39  <andythenorth> am I the only person still playing trunk?
18:23:59  <LordAro> what, svn trunk? :p
18:24:01  <andythenorth> git
18:24:05  <andythenorth> forums are all playing JGR
18:24:09  <andythenorth> pretty much universally
18:24:27  <LordAro> :/
18:24:31  <andythenorth> I assume a chunk of player base is just playing stable release
18:24:48  <LordAro> still seeing plenty of bug reports from 1.8.0
18:24:49  <andythenorth> JGR PP has won
18:25:03  <LordAro> the sooner 1.9.0 can happen the better, imo
18:25:14  <LordAro> also restored nightlies
18:25:16  <andythenorth> well...what can we do about it?
18:25:18  <andythenorth> o_O
18:25:25  <andythenorth> I can either help, or draw pixels
18:25:31  <andythenorth> of give money to Wargaming and play tanks
18:25:43  <andythenorth> of / or /s
18:25:59  <andythenorth> what blocks the new compile farm?
18:26:07  <LordAro> i think nightlies are purely dependent on TB
18:26:09  <nielsm> what needs to be done for nightly builds, apart from possibly something regarding money?
18:26:22  <andythenorth> TrueBrain did you have a trello or something for new CF?
18:26:27  <LordAro> or was it windows builds that were blocking?
18:26:32  <andythenorth> mumble mumble project management
18:26:40  <LordAro> mumble mumble bus factor
18:27:35  <LordAro> it's in my logs somewhere, lemme find it
18:28:28  <LordAro>
18:28:55  <andythenorth> cool
18:29:16  <andythenorth> ok so 'where to host it' is kind of a blocker?
18:29:37  <LordAro> that does sound familiar, because windows needs to be a separate host
18:30:41  <LordAro> i'm gonna do it
18:30:43  <LordAro> @seen TrueBrain
18:30:43  <DorpsGek> LordAro: TrueBrain was last seen in #openttd 1 week, 4 days, 0 hours, 2 minutes, and 26 seconds ago: <TrueBrain> typing is hard
18:30:51  <nielsm> if any junk machine will do I now have a small machine left over
18:30:59  <LordAro> huh, thought it was longer
18:31:07  <nielsm> my old htpc had its hdmi output fried so can't use it for multimedia purposes
18:31:14  <nielsm> but it could still serve as a build host
18:31:15  <nielsm> >_>
18:32:06  <LordAro> iirc TB had a grand plan involving kubernetes
18:32:20  *** Maarten has quit IRC
18:32:25  *** Maarten has joined #openttd
18:32:45  <andythenorth> I did bureacracy
18:33:18  <LordAro> ono
18:33:54  <LordAro> oh, did github-dorpsgek get dumped on the other side of a netsplit?
18:34:01  <Samu> speaking of AIMap.GetMapSizeX()
18:34:11  <Samu> this thing is used almost everywhere
18:34:21  <Samu> maybe i should make it a global variable?
18:34:30  <Samu> or whatever you call it
18:34:31  <LordAro> Samu: well it's not going to change
18:34:48  <LordAro> so at least some sort of constant somewhere would seem appropriate
18:34:54  <LordAro> still probably premature optimisation
18:34:58  <Samu> ah, yes, a constant
18:35:23  <nielsm> if you need the value a lot just pull it into a local
18:36:13  <Samu> before the constructor thing, right?
18:36:16  <nielsm> having it in a global constant within a "namespace class" probably costs nearly just as much as having it behind a function
18:36:31  <Samu> oh :(
18:36:57  <nielsm> at least if it's similar to lua, then the lookup of a member in a global space is a bunch of ops in itself
18:37:52  <nielsm> get global namespace, look up named global, look up field in found object
18:37:59  <nielsm> and call it if it's a function
18:38:07  <nielsm> that's 3 or 4 ops to get
18:39:22  *** Wormnest has quit IRC
18:40:53  <LordAro> andythenorth: well, at least it will mean that TB gets sent an email :p
18:41:48  <Samu> testing
18:41:49  <andythenorth> :)
18:42:39  <Samu> 85
18:42:42  <Samu> yay
18:54:29  <Samu> oh yeah 84 now
18:54:57  <Samu> 	local map_size_x = this._map_size_x;
18:55:07  <Samu> very redundant but saves one more ops
18:59:15  *** tneo has quit IRC
18:59:21  *** tneo has joined #openttd
19:00:03  *** ^Spike^ has quit IRC
19:00:06  <Samu> testing
19:00:08  <Samu> function Road::_IsSlopedRoad(start, middle, end, map_size_x = Road._map_size_x)
19:00:14  *** ^Spike^ has joined #openttd
19:00:30  <Samu> 83!
19:01:02  <Samu> very funny stuff
19:01:26  <LordAro> oho, that's sneaky
19:01:39  <LordAro> remember that code readability is important
19:01:49  <LordAro> making microoptimisations is likely not worth it in the long run
19:02:07  <LordAro> (and it's probably just moving the op to the global level, rather than actually removing it)
19:02:59  <Eddi|zuHause> that sounds like a great deal!
19:04:58  <Eddi|zuHause> "Give us all your data and your firstborn, so we can spam the fuck out of you. you get the chance to win a new smartphone"
19:09:50  <nielsm> <- Samu try this, not sure if it's correct
19:10:33  <nielsm> uh fixed:
19:13:07  <nielsm>
19:13:55  <nielsm>
19:14:45  <Eddi|zuHause> what kind of algorithm is generating your filenames?
19:15:06  <LordAro> sequential with other people using it, i think
19:15:07  *** Gustavo6056 has joined #openttd
19:15:15  <nielsm> yeah, it's not my hosting service
19:15:26  <nielsm> so just sequential in some high base
19:15:47  <Eddi|zuHause> but then, how is p->f->V->W a sequence?
19:16:12  <LordAro> huh
19:16:27  <LordAro> can you delete things, perhaps?
19:16:40  <nielsm> it shouldn't reuse names afaik
19:18:58  <nielsm> if this code is correct and my method of counting is correct too, I think it should top out at 20 ops
19:22:02  *** Gustavo6046 has quit IRC
19:22:02  *** Gustavo6056 is now known as Gustavo6046
19:22:33  *** glx has joined #openttd
19:22:33  *** ChanServ sets mode: +v glx
19:33:03  <LordAro> glx: hey, you're someone who can review & merge things as well :>
19:33:20  <glx> hmm yes
19:33:44  <andythenorth> north-fork
19:33:53  <LordAro> that said there are only 6 that aren't waiting on the author
19:33:59  <andythenorth> yes
19:34:06  <andythenorth> it's actually in pretty good shape
19:34:34  <andythenorth> 151 closed PRs
19:34:47  <LordAro> andythenorth: ok, i'm on your side now - do we close the ones that have been 3 months without activity?
19:35:07  <andythenorth> maybe
19:35:13  <andythenorth> like peter1138's?
19:35:16  <andythenorth> o_O
19:35:23  <LordAro> well if it's been 3 months...
19:35:26  <LordAro> they can always be reopened
19:35:31  <andythenorth> eh we could fix some issues
19:35:36  <andythenorth> here's a great one
19:35:47  <LordAro> haha
19:35:52  <LordAro> shit, yeah
19:36:17  <LordAro> well, we know exactly why it's happening
19:36:21  <LordAro> the issue is how to fix it
19:36:25  <LordAro> i feel like i've said that already
19:38:07  <glx> hmm #6913 probably won't happen
19:38:19  <LordAro> no, i didn't think so
19:38:35  <andythenorth> can't see it
19:38:51  <glx> it's an issue
19:38:52  <andythenorth> too specific to DC?
19:38:59  <andythenorth> I mean I can't see it happening :)
19:39:52  <glx> ha yes too specific, and if we start to add something like that all other similar stuff will be asked
19:40:15  <LordAro> i think that's a "you're welcome to implement it yourself, but lol no"
19:40:48  <Eddi|zuHause> that sounds like something for an admin port library?
19:40:57  <glx> next they will ask for twitch integration
19:41:31  <Eddi|zuHause> same response
19:41:43  <glx> ah yes I think an admin port library could do that
19:42:28  <nielsm> "run this other program too to make the game show up" is not a great solution though
19:43:07  <LordAro> some sort of "ottd_discord" wrapper wouldn't be too difficult from there though
19:44:09  <nielsm> otherwise a kind of plugin system
19:44:23  <andythenorth> all your openttd belong to us
19:48:13  <LordAro> i could hack it so that if you've got "Apple LLVM" compiler, it just forces it to some reasonable version
19:49:26  <LordAro> or we could revert that change so that c++11 isn't actually "required"
19:49:28  <LordAro> or both
19:50:43  <Eddi|zuHause> what is the actual problem? it can't read the version number correctly?
19:50:49  <LordAro> basically, yeah
19:51:01  <LordAro> because apple are stupid and wipe out all trace of the actual clang version number
19:51:05  <Eddi|zuHause> does it maybe choke on some unprintable characters like colour codes?
19:51:35  <andythenorth> it's something like this (or similar)
19:52:28  <Eddi|zuHause> brb
19:53:19  *** Eddi|zuHause has quit IRC
19:53:57  *** nielsm has quit IRC
19:55:18  *** sla_ro|master has quit IRC
19:56:12  *** Gja has quit IRC
20:14:59  *** Gustavo6056 has joined #openttd
20:21:35  *** Thedarkb-T60 has joined #openttd
20:22:33  *** Gustavo6046 has quit IRC
20:22:33  *** Gustavo6056 is now known as Gustavo6046
20:24:09  *** Thedarkb1-T60 has joined #openttd
20:25:03  *** Eddi|zuHause has joined #openttd
20:26:31  <LordAro> andythenorth: can you add a `echo "foo: $cc_version"` on line 1329 of config.lib ?
20:27:01  <LordAro> according to the issue, you should be getting "version" 91
20:27:13  <LordAro> issue comments*
20:27:19  <LordAro> but that shouldn't be causing any issues#
20:27:23  <andythenorth> ok BRB
20:28:28  <LordAro> if [ "$cc_version" -lt "30" ]; then branch is triggering, but i don't see how
20:30:24  *** Thedarkb-T60 has quit IRC
20:31:52  <andythenorth> foo: 10
20:32:01  <andythenorth> printed twice
20:32:44  <LordAro> hmm
20:32:48  <andythenorth>
20:32:50  <LordAro> well that explains the version issue
20:33:19  *** Eddi|zuHause has quit IRC
20:33:21  <LordAro> for my own peace of mind, can you run `cc -v 2>&1 | head -n1` ?
20:33:35  <LordAro> and also clang -v 2>&1 | head -n1
20:33:36  <andythenorth> Apple LLVM version 9.1.0 (clang-902.0.39.2)
20:33:45  <andythenorth> Apple LLVM version 9.1.0 (clang-902.0.39.2)
20:33:52  <LordAro> gcc ?
20:34:42  <andythenorth> clang: error: no input files
20:34:56  <LordAro> `gcc -v 2>&1 | head -n1` :p
20:35:02  <andythenorth> I wondered :P
20:35:10  <andythenorth> I learnt to only paste what people type :P
20:35:23  <andythenorth> Configured with: --prefix=/Applications/ --with-gxx-include-dir=/Applications/
20:35:46  <LordAro> that's... really stupid
20:35:51  <LordAro> but explains where the 10 comes from
20:36:00  <LordAro> (MacOSX10.13.sdk)
20:36:04  <andythenorth> yair
20:36:45  <andythenorth>
20:37:19  *** Eddi|zuHause has joined #openttd
20:37:51  <andythenorth> Eddi|zuHause: so why won't this work then? o_O
20:37:57  * andythenorth assumes there's a reason
20:39:12  <Eddi|zuHause> andythenorth: it was never implemented because... it's complicated (tm)
20:39:24  <LordAro> hmm, i guess that could work... replace 1328 with: cc_version="` -v 2>&1 | grep -i version | head -n 1 | sed s@[^0-9]@@g | cut -c 1-2`"
20:40:01  <andythenorth> foo: 91
20:40:10  <LordAro> excellent
20:40:13  <LordAro> try compiling
20:40:34  * andythenorth compiling
20:41:40  <andythenorth> not sure if I have the vars set here, do they persist past a computer restart?
20:41:41  <andythenorth>
20:41:51  <andythenorth> or I could just explicitly clear them
20:42:04  <andythenorth> oh meanwhile
20:42:05  <andythenorth> clang: error: linker command failed with exit code 1 (use -v to see invocation)
20:42:06  <LordAro> andythenorth: not unless you `export`ed them
20:42:17  <LordAro> that's disappointing
20:42:25  <Eddi|zuHause> andythenorth: anyway, CB36 doesn't have any way to access the to-be-shuffled properties, because of recursiveness
20:42:27  <LordAro> that could be the iconv thing again
20:42:40  <andythenorth> Eddi|zuHause: that's kind of a blocker :)
20:42:53  <Eddi|zuHause> andythenorth: proper way is to actually reorder the articulated vehicles
20:42:54  <LordAro> andythenorth: just to confirm, LDFLAGS="-liconv" ./configure && make ?
20:43:18  <andythenorth> LordAro: works now
20:43:20  <andythenorth> with that
20:43:37  <LordAro> mm
20:43:45  <andythenorth> Eddi|zuHause: how will I do stupid livery tricks on flip then? :P
20:43:52  <andythenorth> if the vehicles actually reverse :(
20:43:58  <Eddi|zuHause> andythenorth: you don't.
20:44:50  <andythenorth> LordAro: so does that advance us any further? o_O
20:45:30  <Eddi|zuHause> more important question: try to resurrect this starting-to-show-old-age computer or just suck it and buy a new one?
20:45:39  <LordAro> andythenorth: it does!
20:46:01  <andythenorth> Eddi|zuHause: try to resurrect it
20:46:07  <andythenorth> then regret
20:46:18  <Eddi|zuHause> already beyond those two phases
20:46:20  <LordAro> next issue is iconv(.h) detection, but it looks like you can override that for now with ./configure --with-iconv - no need for environment variables
20:46:55  <LordAro> if you want to continue, i'd be interested in the output of `find / -name iconv.h 2>/dev/null`
20:46:55  <andythenorth> configure: error: iconv couldn't be found
20:47:00  <LordAro> damn
20:47:07  <andythenorth> that wasn't output from find
20:47:16  <andythenorth> that was from configure parameter
20:47:23  <Eddi|zuHause> --with-iconv=path?
20:47:28  <LordAro> ah yeah, i misread the thing
20:47:36  <LordAro> Eddi|zuHause: that'd work, just a matter of determining the path
20:47:45  <LordAro> and working out why the configure script isn't checking it by default
20:47:54  <andythenorth> the find command prints the path
20:48:01  <andythenorth> "/usr/local/Cellar/libiconv/1.15/include/iconv.h"
20:48:03  *** Exec has quit IRC
20:48:03  *** Exec has joined #openttd
20:48:13  <LordAro> just that one?
20:48:17  <andythenorth> just that one
20:48:26  <andythenorth> obvs. varies per mac eh
20:48:48  <LordAro> yeah...
20:56:17  *** Thedarkb1-T60 has quit IRC
20:56:32  *** Thedarkb-T60 has joined #openttd
20:57:33  *** Progman has quit IRC
20:58:13  <LordAro> andythenorth: at the risk of asking the obvious, are you sure iconv is installed "properly" ?
20:58:31  <andythenorth> it's installed using one of the 2 sensible ways
20:58:58  <andythenorth> I could reinstall
20:59:46  <andythenorth> I'll see if there's any upgrade
21:00:02  * andythenorth waiting for ports tree :P
21:03:31  <LordAro> for whatever reason, the build system is expecting /usr/include/iconv.h or /usr/local/include/iconv.h
21:03:47  <LordAro> i don't know enough about OSX to know why you've got a /usr/local/Cellar
21:03:55  <LordAro> and why iconv is in there
21:04:15  <LordAro> but fwiw, ./configure --with-iconv=/usr/local/Cellar/libiconv/1.15 will work
21:04:38  <lethosor_> That's from homebrew
21:04:55  <andythenorth> the options on macos are homebrew or macports
21:05:03  <andythenorth> or installing your own in the system
21:05:16  <andythenorth> which inevitably breaks when Apple upgrade the OS
21:05:27  <lethosor_> Generally Homebrew will install packages to /usr/local, but maybe not for iconv
21:06:26  <LordAro> hmm, i could do some fanciness to just iterate over the searchpaths...
21:06:44  <LordAro> instead of just /usr/include & /usr/local/include
21:08:04  <andythenorth> allegedly homebrew is supposed to symlink /usr/local/opt/
21:09:31  <andythenorth> seems it does afaict
21:12:22  *** Eddi|zuHause has quit IRC
21:12:39  *** Eddi|zuHause has joined #openttd
21:13:03  <lethosor_> LordAro: homebrew normally links packages it installs into /usr/local. In the case of libiconv, it does not, because macOS distributes libiconv itself
21:13:19  <andythenorth> such fun
21:13:23  <lethosor_> brew info libiconv: "Caveats: libiconv is keg-only, which means it was not symlinked into /usr/local, because macOS already provides this software and installing another version in parallel can cause all kinds of trouble."
21:13:58  <lethosor_> are you still having issues with linking it? not that it helps, but it works fine for me with homebrew, despite that
21:15:05  *** Samu has quit IRC
21:15:15  *** Gustavo6056 has joined #openttd
21:15:59  <LordAro> nothing has any issue linking it
21:16:10  <LordAro> all andythenorth needs to do is force it with ldflags
21:16:33  <LordAro> but the build system is only checking /usr/{local/,}include/ for iconv.h
21:16:42  <LordAro> hang on, i'm putting a branch together
21:17:38  *** Samu has joined #openttd
21:17:40  <Samu> back
21:18:44  *** ttech2 has quit IRC
21:19:47  *** Ttech has joined #openttd
21:19:47  <LordAro> andythenorth: feel like trying ?
21:20:19  <Samu> where is nielsm
21:21:48  <andythenorth> probably gone to bed
21:21:57  *** rocky113844 has quit IRC
21:22:32  *** Gustavo6046 has quit IRC
21:22:32  *** Gustavo6056 is now known as Gustavo6046
21:22:39  * andythenorth cloning
21:25:54  <Samu> got 49 ops
21:26:02  <Samu> he said it would top at 20
21:26:04  <Samu> :(
21:28:04  *** frosch123 has quit IRC
21:28:07  <LordAro> Samu: well that's still nearly half what you had
21:30:01  <Samu> wait,
21:30:03  <Samu> 44
21:30:18  <Samu> im not sure what is to edit out
21:30:32  <andythenorth> LordAro: test compiling :P
21:30:38  <LordAro> \o/
21:30:51  <andythenorth> eh, let's see if it finished
21:31:52  <Samu> does his code account for IsSteepSlope
21:31:53  <Samu> ?
21:32:06  <LordAro> i think that's for you to work out
21:32:29  <andythenorth> LordAro: compiled, runs
21:32:32  <andythenorth> :)
21:32:36  <LordAro> \o/
21:32:38  * LordAro PRs
21:32:42  <andythenorth> awesome thanks
21:32:48  <Samu> too cody for my brain
21:33:03  *** HerzogDeXtEr has quit IRC
21:34:42  <Samu> i'm always getting 44, that seems fishy
21:36:09  <Samu> i have an idea
21:36:50  <Samu> to see if his function outputs the same result, i'm gonna use this function twice and then assert compare between each other
21:37:08  *** rocky1138 has quit IRC
21:37:31  *** rocky1138 has joined #openttd
21:41:18  <LordAro> wait, crap, i can create branches on OpenTTD/OpenTTD as well?
21:41:33  <andythenorth> :o
21:41:37  <andythenorth> the power
21:41:38  * LordAro deletes
21:41:40  <LordAro> you saw nothing
21:41:57  <LordAro> still can't merge PRs though :p
21:42:53  <Samu> assert(self._IsSlopedRoad(par_tile, prev_tile, new_tile) == self._IsSlopedRoadEfficient(par_tile, prev_tile, new_tile));
21:43:01  <Samu> keks
21:43:57  *** gelignite has quit IRC
21:44:57  <Samu> assertion failed... rip
21:45:27  <Samu> either i misunderstand what needs to be edited, or...
21:59:29  <Samu> local t = AITile; this didn't save any op
21:59:53  <Samu> remained the same
22:01:10  <glx> LordAro: searchpath looks like brute force ;)
22:01:26  <LordAro> yup!
22:01:42  <Samu> nevermind, i was edited the wrong file, grr
22:01:51  <glx> OSX is worse than MSYS/MSYS2
22:02:44  <lethosor_> glx: what?
22:03:15  <glx> we needed to add many special cases for MSYS2
22:03:40  <glx> but not like that
22:03:59  <lethosor_> Checking /usr/local/Cellar is too much work?
22:04:08  <Samu> actually it helps
22:04:13  <Samu> 76 ops
22:05:00  <glx> lethosor_:
22:06:09  <lethosor_> Looks like more than necessary to me
22:06:36  <glx> I think it's for the next time something changes in OSX world ;)
22:07:04  <lethosor_> It's not a recent change by any means
22:07:22  <lethosor_> Homebrew has never installed iconv system-wide
22:09:28  <LordAro> i don't understand how it became an issue recently though
22:09:45  <LordAro> did older osx just leave an old iconv.h header lying around or something?
22:10:17  *** quiznilo has joined #openttd
22:10:31  <lethosor_> It's a system-specific issue somehow. It works fine for me, like I said
22:13:53  <andythenorth> it only started happening on a clean install of macOS for me
22:14:02  <andythenorth> old one probably cargo culted something
22:14:15  *** rocky113844 has joined #openttd
22:15:18  *** Gustavo6056 has joined #openttd
22:15:29  *** quiznilo has left #openttd
22:19:38  <Samu> 46/44/77/77/77/46/44/75/71/77/72 - _IsSlopedRoadEfficient
22:21:25  <LordAro> (wrote a load more text in #6917 for future people)
22:21:46  <Samu> 70/68/107/107/107/68/66/105/97/107/103 - _IsSlopedRoad (the original code)
22:22:32  *** Gustavo6046 has quit IRC
22:22:32  *** Gustavo6056 is now known as Gustavo6046
22:23:17  <Samu> can't manage to make nielsm code to not assert
22:24:45  <Samu> here's both for comparison
22:24:56  <Samu> what I was able to do without asserting
22:25:38  *** andythenorth has left #openttd
22:32:20  *** HeyCitizen has quit IRC
22:33:26  *** sushibear_ has quit IRC
22:33:57  *** HeyCitizen has joined #openttd
22:34:12  *** rocky113844 has quit IRC
22:41:21  *** rocky113844 has joined #openttd
22:45:07  <Eddi|zuHause> <LordAro> wait, crap, i can create branches on OpenTTD/OpenTTD as well? <- sounds like something that should be prevented?
22:51:04  <LordAro> probably
22:51:39  *** Thedarkb-T60 has quit IRC
22:51:59  *** Thedarkb-T60 has joined #openttd
23:01:34  *** Thedarkb-T60 has quit IRC
23:15:30  *** Gustavo6056 has joined #openttd
23:17:17  *** haudrauf has quit IRC
23:21:30  *** haudrauf has joined #openttd
23:22:32  *** Gustavo6046 has quit IRC
23:22:32  *** Gustavo6056 is now known as Gustavo6046
23:29:54  *** haudrauf has quit IRC
23:31:37  *** haudrauf has joined #openttd
23:45:08  *** Extrems has quit IRC
23:45:15  *** Extrems has joined #openttd
23:45:28  *** Samu has quit IRC

Powered by YARRSTE version: svn-trunk