IRC logs for #openttd on OFTC at 2009-10-19
⏴ go to previous day
00:00:45 <fonsinchen1> OK, I'll get some sleep now. Good night
00:18:31 <mynetdude> is there any way to upgrade the existing airport without tearing it down?
00:18:56 <mynetdude> like build on top of it? I wan to build a larger airport but to do that would mean to build it on the other side of the city
00:19:24 <mynetdude> I want to keep it connected to any rail stations as well
00:25:12 <Sitethief> just demolish the current one and build a new one? whats the problem?
00:26:00 <mynetdude> it doesn't say how much it is... and if my city will not permit a new one to be built then I'm screwed
00:26:15 <Sitethief> press shift while building
00:26:41 <Sitethief> and you can check the city status, if its anything above medirocre (iirc) it won't be a problem
00:27:39 <mynetdude> what does the shift do while buidling?
00:27:48 <Sitethief> shift shows you buildign cost without building it
00:31:15 *** Sitethief is now known as Sitethief{afk}
00:36:04 <mynetdude> just have to do it when nothing is around and reroute planes to the new field
00:39:32 <Sitethief{afk}> why reroute them? place the new airfield on top of the old and its will keep the same name, no planes need to be rerouted
00:47:13 *** Rubix`` has joined #openttd
00:56:54 *** thingwath has joined #openttd
01:03:10 <mynetdude> Sitethief{afk} because you can't build on top of buildings?
01:03:30 <Sitethief{afk}> demolish the old one,a build the new one at the same spot, its gonna have the same name
01:03:32 <mynetdude> that is what I tried to do, but you have to demolish the old one first?
01:07:12 <mynetdude> well I got bored of the game, saved it for later
01:09:00 *** KenjiE20|LT has joined #openttd
01:13:03 <CIA-9> OpenTTD: rubidium * r17812 /trunk/src/ (cargopacket.cpp cargopacket.h): -Codechange: move the feeder_share cache from CargoList to VehicleCargoList; saves 512 bytes per station and 1-2% on CargoList::MoveTo.
01:13:47 *** thepalm has joined #openttd
01:20:05 <break19> I wish you could "expand airports" that way.. build a new on top.. If the airport is extremely busy, it may be damn near impossible to demolish the old one, as there will always be a plane at it..
01:20:33 <Rubidium> well... airports can't be expanded anyway when there's a plane on it
01:20:49 <Rubidium> so it wouldn't solve your 'there needs to be no plane on it' problem at all
01:21:33 <break19> Rubidium: you're quite right, obviously, but there needs to be a way to put airports into "no landings" mode, happens all the time in the real world..
01:23:40 <break19> add an element to the airfield, possibly at the tower, it pops up a window with a "Service Airfield" button. click it, the airfield can no longer accept landings, any aircraft currently loading can finish loading and takeoff.
01:24:01 <break19> then, when the airport is empty, demo/rebuild/etc
01:27:12 <Sitethief{afk}> just link all the plane routing together, send them all to one airfield, demolish, build new, reinstae old route, done
01:28:07 *** thepalm has joined #openttd
02:01:42 <Pikka> <Rubidium> well... airports can't be expanded anyway when there's a plane on it <- they could be according to my spec :P
04:39:34 *** lobster has joined #openttd
05:35:19 *** Rhamphoryncus has joined #openttd
05:43:22 *** Dred_furst has joined #openttd
06:10:05 *** welshdragon has joined #openttd
06:11:52 *** Cybertinus has joined #openttd
06:43:01 *** Polygon has joined #openttd
07:02:35 *** welshdragon has joined #openttd
07:02:50 *** Terkhen has joined #openttd
07:24:12 *** boekabart has joined #openttd
07:29:08 *** ChanServ sets mode: +v tokai
07:46:49 *** Sitethief{afk} has quit IRC
08:00:45 *** ChanServ sets mode: +v tokai
08:14:04 <boekabart> only for some subfolders, like src\video\cocoa
08:21:56 <boekabart> happens when svn upping from 11127 to 11500
08:26:49 <TrueBrain> boekabart: remove those subfolders and run svn update again?
08:27:32 <boekabart> same happens when snv upping for HEAD to 11127
08:28:29 <TrueBrain> the problem is like this: you send secure.openttd.org, it is reached by a proxy, it is send to the Apache2 WebDAV, which handles the request
08:28:57 <TrueBrain> now this is not the problem of course, but it is this: all URLs are rewritten to make the Apache2 WebDAV understand what is going on (it is not running on port 80, and else he doesn't understand the request)
08:29:01 <boekabart> https specific problem?
08:29:23 <TrueBrain> certain SVN WebDAV aspects have a secondary link, like 'COPY' and 'MOVE'. This is not rewritten. Here is where Apache2 freaks out and aborts the connection
08:29:33 <TrueBrain> (in results your client gives a bogus error)
08:29:53 <TrueBrain> we tried rewritting the url, it fails .. Apache2 wants something which is impossible to give to him
08:30:08 <TrueBrain> short: Apache2 WebDAV is broken to be used AFTER a proxy
08:30:15 <TrueBrain> so http/https specific, yes
08:30:27 <TrueBrain> what mostly helps for me is to checkout a complete new version
08:30:34 <boekabart> yes, that does help
08:30:35 <TrueBrain> it only happens with 'moves' and 'copies'
08:30:54 <boekabart> but i'm trying to 'upgrade' a patch
08:31:06 <TrueBrain> I understand .. temporary use 'svn://'? :)
08:31:26 <TrueBrain> you don't have a ssh account :p
08:31:28 <boekabart> svn not due to network limitations
08:33:01 <TrueBrain> the NoAI projects notice this problem the most, as they can only access their repos over http/https. I don't really have a solution ...
08:33:18 <TrueBrain> svn diff > temp.patch && svn -R revert . && svn update && patch -p0 < temp.patch :p
08:33:23 *** boekabart_ has joined #openttd
08:33:23 *** boekabart is now known as Guest1598
08:33:23 *** boekabart_ is now known as boekabart
08:34:04 <TrueBrain> oh, and additional to the above problem, internally nothing is https, so Apache also wines that the incoming connection is a http, and the 2nd url in the header is https :p
08:34:27 <boekabart> so http might work better? In that case I could try a proxy
08:34:37 <TrueBrain> you can try, but last time it gave problems too
08:37:06 <boekabart> Ammler: does hg up do merging as well as svn? (better is also OK)
08:38:34 <Ammler> he, don't have real experience to make a statement :-)
08:43:03 <TrueBrain> boekabart: does it work now?
08:44:14 <TrueBrain> Provider encountered an error while streaming a REPORT response. [500, #0]
08:44:18 <TrueBrain> I love 'clear' errors :)
08:44:35 <boekabart> i'll be back in an hour
08:45:03 <boekabart> hey, http is OK by the way, it seems
08:50:07 <TrueBrain> then that will be our solution for now, as it fails on REPORT, which is very weird
08:50:14 <TrueBrain> and I have to go .. maybe I will look at it another day :)
08:51:10 *** HerzogDeXtEr1 has joined #openttd
09:08:14 *** Chris_Booth has joined #openttd
09:11:51 *** Biolunar has joined #openttd
09:15:59 <CIA-9> OpenTTD: rubidium * r17813 /trunk/src/ (4 files in 2 dirs): -Codechange: unify the CargoPacket related coding style
09:26:34 *** fonsinchen has joined #openttd
09:34:18 <fonsinchen> Rubidium: Are you going to accept my proposal about different containers for station and vehicle cargo lists? If yes, you're going to run into problems if you use the CargoList instances as template parameter. You can't use a template parameter's typedefs in the templated class in C++. But maybe I'm wrong and you have a different idea there. Could you elaborate some?
09:37:35 <fonsinchen> OK, that works because it has a second tparam. I still don't see the necessity, but I can live with that.
09:41:38 *** welshdragon has left #openttd
09:55:35 <Rubidium> did you know that 50% of CPU time of MoveTo is spent in splitting cargo packets?
09:59:43 <Rubidium> fonsinchen: for the 'moving' of feeder_share the Tinst is needed
10:02:49 <fonsinchen> Why don't you make RemoveFromCache and AddToCache virtual? Then you don't need the second tparam. And why is the splitting so expensive? Because of the memory allocation? Shouldn't the pool handle that?
10:04:02 <Rubidium> making RemoveFromCache/AddToCache virtual made it slower :(
10:04:45 <Rubidium> the new pool does malloc/free, which isn't bad in general but for cargopackets that are quite short lived it might be troublesome
10:05:15 <fonsinchen> Ah, I thought it had a buffer and did placement new
10:05:46 <Rubidium> on the other hand, improving the 2000 CPU cycles to 20 CPU cycle would yield 1ms per 2 seconds
10:06:01 <Rubidium> which is basically negligable
10:06:09 <fonsinchen> This might be a better idea in case of cargopackets. Also the zeroing can be dropped as all the members are initialized by the constructor anyway.
10:09:22 <fonsinchen> How bad is the effect of virtual AddTo/RemoveFromCache? Basically it has to do vtable lookups, but the class hierarchy is rather shallow so that shouldn't take too long.
10:09:27 *** Belugas has joined #openttd
10:09:28 *** ChanServ sets mode: +o Belugas
10:13:46 <fonsinchen> And of course the time spent in (the old) Append, also when called from MoveTo, depends on the number of different cargopackets in the list. With cargodist and at stations with lots of transfers that number grows.
10:14:47 *** ChanServ sets mode: +v tokai
10:23:08 <Rubidium> with static_cast is MoveTo is 2-3% faster
10:29:26 *** worldemar has joined #openttd
10:32:57 <Rubidium> Append (with non-virtual) is 2-8% faster than Append with virtual functions
10:34:40 <fonsinchen> The question is if that's a station with a single cargopacket (like 90% of the stations without cargodist) or if it's a station with transfers.
10:36:34 <Rubidium> it's quite an old savegame with a few transfers, but lots of transported cargo
10:38:35 <Rubidium> so caching is quite a big portion of the relative time spent
10:38:37 <fonsinchen> The number of cargopackets only explodes if you have a lot of transfers. Otherwise there is only the cargo generated at the station itself and that is always merged into a single packet. When loading onto a vehicle there is typically also only one packet then.
10:41:23 <Rubidium> so... ottdcoop people... what's the latest *massive* trunk savegame of you that used transfers quite a lot?
10:48:31 *** TheMask96 has joined #openttd
11:03:40 <fonsinchen> Public Server #154 has some transfers
11:03:51 <fonsinchen> and it's horribly slow
11:04:06 <Rubidium> ottdcoop prozone5 gives me: 17% speed increase in append over virtual AddCache/RemoveCache and 2% for MoveTo
11:05:30 <fonsinchen> So, then let's stop discussing the virtual Add/RemoveCache. You are right there.
11:06:23 <Rubidium> hmm.. ps154 does way more :)
11:06:40 <fonsinchen> I will come up with it again, when the overall number of packets has been reduced by changing the containers.
11:07:07 <fonsinchen> s/packets/packets that need to be checked in Append/
11:07:13 *** andythenorth has joined #openttd
11:07:34 <Eddi|zuHause> isn't the template vs virtual part the entire reason why YAPF is way faster than NPF?
11:07:54 <Rubidium> but reducing the number of packets to check implies that the relative weight of the cache updates gets higher
11:09:00 *** Biolunar has joined #openttd
11:14:24 <Rubidium> for PS154: 2% faster MoveTo, 5% faster Append
11:20:26 *** andythenorth has joined #openttd
11:30:51 *** Chruker has joined #openttd
11:41:47 *** KenjiE20 has joined #openttd
11:42:22 *** Chillosophy has joined #openttd
12:25:49 *** andythenorth has joined #openttd
12:47:15 * boekabart misses a 'rating' system in the in-game content download system
12:47:21 <boekabart> ... for AI's mainly
12:48:26 <Eddi|zuHause> yeah, indicators for maturity level and difficulty level...
12:56:00 <boekabart> I suppose it's up to the creator/publisher whether a certain newgrf is in there or not?
12:56:13 <boekabart> like dutch trains, catenary and stuff missing, author's choice?
12:56:46 <SpComb> one shouldn't upload stuff that isn't ones own, no
12:56:51 <Eddi|zuHause> the terms of use say, only an author can upload stuff
12:56:54 <Ammler> the download statistic isn't very helpful
12:57:08 <SpComb> perhaps the author hasn't been bothered to, perhaps they don't want to
12:57:30 <Eddi|zuHause> i don't know who's the author of dutch catenary
12:57:40 <Eddi|zuHause> maybe he doesn't even know?
12:57:47 *** Brianetta has joined #openttd
12:57:48 <Ammler> boekabart: just ask in the threads, if there are enough requests, maybe they do :-)
12:58:42 <Ammler> Eddi|zuHause: www.openttdcoop.org/wiki/GRF_Table would tell you ;-)
12:58:56 <Eddi|zuHause> why would i bother finding out?
12:59:20 <Brianetta> Xeryus, are you here?
12:59:27 <Eddi|zuHause> i already have the file... i'm not really interested in getting it...
12:59:56 *** duckedtapedemon has quit IRC
13:00:08 <Rubidium> how would download statistics ever be useful?
13:00:43 <Rubidium> Brianetta: given that he's disconnected from his bouncer, very slim chance
13:01:59 <Brianetta> I know he's on a train right now, I was hoping he was using the free wifi
13:02:33 <SpComb> call him and tell him!
13:02:46 <Brianetta> Since calling or texting his Dutch mobile is really expensive
13:03:39 <Rubidium> and the 'fun' is, it's equally expensive for him to receive it
13:04:26 <Brianetta> He should pick us a UK sim
13:04:53 <Rubidium> how long is he going to be in the UK and is he going to call/be called
13:05:46 <Rubidium> no idea how expensive and how much trouble it is to get a sim in the UK
13:07:52 <Rubidium> e.g. for Japan you need to be a resident to be able to buy a phone with prepaid sim (they don't sell them separately)
13:09:07 <boekabart> Download stats aren't useful, but download rating might be
13:09:19 <boekabart> s/download rating/user rating
13:09:38 * boekabart misses skype's ctrl-up to repair a typo
13:09:48 <boekabart> irc feature request :)
13:10:20 * boekabart also concludes that updating his deepwater patch is virtually impossible - close to a remake
13:10:40 <boekabart> devs, pls stop changing the code so much!
13:11:59 * Belugas keeps on obeying to Sir boekabart
13:13:21 <boekabart> Belugas: don't abuse my remark as an excuse to be lazy
13:19:12 *** Dred_furst has joined #openttd
13:23:18 <boekabart> your last commit, 10 weeks old :)
13:23:29 <boekabart> that's older than my son :D
13:33:47 <Eddi|zuHause> then you sound like you have all night to code ;)
13:34:14 *** Coco-Banana-Man has joined #openttd
13:46:50 <boekabart> Eddi|zuHause: me? i wish
13:47:30 *** `Fuco`OFF has joined #openttd
13:49:59 <Belugas> someone sends out his wife when baby is crying...
13:52:34 *** `Fuco`OFF has joined #openttd
13:56:32 *** `Fuco`OFF is now known as FUco
13:57:54 <hickop> is there a trick to increase "transported" value of industries or its random ?
13:58:45 <Eddi|zuHause> the "transported" value corresponds to the average station rating
13:59:14 <Eddi|zuHause> strategies to increase station rating are around buried somewhere in the wiki
13:59:53 <Eddi|zuHause> the easiest ones are "have always a train waiting" and "build a statue in the nearby town"
14:15:16 *** Biolunar has joined #openttd
14:19:45 *** Doorslammer has joined #openttd
14:20:36 <boekabart> water_map.h line 151 MakeWater(...) _really_ should be called MakeSea(...)
14:21:04 <boekabart> IsSea, IsCanal, IsRiver but MakeWater, MakeCanal, MakeRiver
14:22:30 <Belugas> awter been a general description maybe? dunno... could be
14:23:34 <boekabart> nope, look at it: really makes SEA
14:24:06 <boekabart> ' _m[t].m3 = WATER_CLASS_SEA; '
14:25:21 <andythenorth> apparently boats in TTDPatch can have different travel speeds for canals / open sea
14:25:25 <andythenorth> kind of interesting no?
14:25:34 <andythenorth> Or maybe a bit too much realism?
14:33:19 *** Grelouk has joined #openttd
14:36:27 *** Grelouk_ has joined #openttd
14:38:54 *** Grelouk__ has joined #openttd
14:58:25 *** Chillosophy has joined #openttd
14:59:34 *** andythenorth has joined #openttd
15:01:44 *** boekabart has left #openttd
15:09:21 *** TrainzStoffe has joined #openttd
15:15:30 *** Progman has joined #openttd
15:16:19 *** TrainzStoffe is now known as Stoffe
15:24:33 *** insulfrog has joined #openttd
15:28:25 *** ChanServ sets mode: +v tokai
15:36:48 <CIA-9> OpenTTD: rubidium * r17814 /trunk/src/cargopacket.h: -Codechange: there's no need to invalidate the cache in the constructor of a CargoList; the list is empty, the CargoList is calloc-ed so all caches are 0.
15:50:58 *** ChanServ sets mode: +v tokai
16:01:15 *** worldemar has joined #openttd
16:01:18 *** Rhamphoryncus has joined #openttd
16:17:01 *** ChanServ sets mode: +v tokai
16:22:46 *** andythenorth has joined #openttd
16:28:53 *** worldemar has joined #openttd
16:40:16 *** ChanServ sets mode: +v tokai
16:41:00 *** frosch123 has joined #openttd
16:46:32 *** Dreamxtreme has joined #openttd
16:48:10 *** Brianetta has joined #openttd
16:54:28 *** fonsinchen has joined #openttd
16:59:23 *** TrainzStoffe has joined #openttd
17:00:32 *** andythenorth has joined #openttd
17:04:27 *** welshdragon has joined #openttd
17:06:19 *** TrainzStoffe is now known as Stoffe
17:12:01 *** insulfrog has left #openttd
17:18:30 *** TheMask96 has joined #openttd
17:28:38 *** TrainzStoffe has joined #openttd
17:32:50 *** Terkhen has joined #openttd
17:35:37 *** TrainzStoffe is now known as Stoffe
17:55:04 *** StarLionIsaac has joined #openttd
18:24:59 *** andythenorth has joined #openttd
18:33:45 <frosch123> Sacro: stop sending spam to openttd addresses
18:54:55 *** boekabart has joined #openttd
19:05:20 *** andythenorth has joined #openttd
19:05:43 <frosch123> sorry, i have some danish decease
19:06:11 <Sacro> you're from denmark too eh
19:06:30 <DorpsGek> Sacro: I have not seen Bjrani.
19:06:47 <DorpsGek> Sacro: Bjarni was last seen in #openttd 5 weeks, 5 days, 23 hours, 6 minutes, and 42 seconds ago: <Bjarni> Belugas: there is no reason to write to PeterT. We already know he is unable to read anyway :P
19:06:50 <frosch123> you also have some decease, right?
19:07:56 <frosch123> hmm, they are quite close, aren't they?
19:08:49 *** welshdragon has joined #openttd
19:14:50 <DorpsGek> andythenorth: Roujin was last seen in #openttd 5 weeks, 5 days, 1 hour, 31 minutes, and 12 seconds ago: <Roujin> ...and demand everything spoon-fed to them. Anyway, it seems you're not in a helpful mood, so I'll not bother you anymore... :/
19:15:27 *** HerzogDeXtEr has joined #openttd
19:26:36 <Sacro> Pikkaa: whyyyyy a class 58 XD
19:27:01 <Sacro> the only fondness i have for them is due to the HO guage modeli hav
19:27:25 <Pikkaa> what's wrong with a class 58?
19:28:07 <Pikkaa> it's good to have /something/ in there
19:28:11 <Pikkaa> and they're different-looking
19:29:01 <Sacro> alright for general freight i guess
19:29:46 <Sacro> i have that in OO it seems
19:30:41 <andythenorth> what got you two talking about bones?
19:32:01 <Sacro> there's gonna be in UKRS
19:33:37 *** fonsinchen has joined #openttd
19:36:39 <Sacro> TrueBrain: could start specifying 2k/XP/Vista/7 on the download page now
19:38:25 <andythenorth> ^^ that one's a nice colour too
19:41:16 <andythenorth> Belugas is not a 58
19:41:25 <andythenorth> Pikkaa is not a 58
19:41:39 <andythenorth> Pikkaa is being greedy with the letter a today though
19:47:24 *** Pikkaa is now known as Pikk
19:55:05 <andythenorth> Pikka was abstemious with the letter a....briefly
19:56:31 <andythenorth> ^^ ahem. Charts, not graphs. You just made my inner autist wince
19:57:39 <Pikka> yes but my inner navigator knows that a chart is a map
19:59:09 <andythenorth> my inner excel just crashed :|
20:01:04 <andythenorth> your charts might be, but my excel is seriously sulking
20:01:18 <andythenorth> I want charts dammit :x
20:01:21 <Pikka> what did you do to it? :P
20:01:42 <andythenorth> the usual, tried to use the features
20:02:18 * Pikka uses open office, but in no way feels superior about it
20:04:32 <andythenorth> I fancied a chart of cost-per-ton for FISH, but....another day
20:05:18 <Prof_Frink> What's that got to do with the price of fish?
20:09:47 *** Dreamxtreme has joined #openttd
20:14:18 <welshdragon> what's that got to do with the price of your mum, Prof_Frink?
20:22:22 <Dreamxtreme> price of your mum ?
20:26:39 <Belugas> for anythung else, there's credit card
20:27:27 <welshdragon> that fails so badly, Belugas
20:30:27 *** Progman has joined #openttd
20:30:58 <andythenorth> I have charts too! lookit me, lookit me!
20:32:20 <CIA-9> OpenTTD: rubidium * r17815 /trunk/src/video/sdl_v.cpp: -Fix [SDL]: asynchronious drawing caused extra unresponsiveness during map generation; disable the threading while generating a map
20:34:21 <andythenorth> right time for nfo
20:34:37 <andythenorth> OMG, I am responding to user requests for features :O
20:35:30 *** fonsinchen has joined #openttd
20:36:10 <andythenorth> wish there was a varaction 2 for ships to see if we are sailing at sea, or in a canal
20:37:04 <andythenorth> and also a cb for 'sink this boat'
20:38:34 <andythenorth> "Boat number 32 sailed too far out to sea and sank"
20:39:15 <frosch123> you want to add triremes to fish?
20:39:46 <andythenorth> I want to add a refit option "Overload dangerously"
20:40:03 <Belugas> welshdragon : looks like you have been too much intoxicated by the adds ;)
20:40:28 <welshdragon> no, the fact that it's a Mastercard ;)
20:40:30 <boekabart> Rubidium: did you see my remark about 'MakeWater' ?
20:40:42 <andythenorth> cb 15...refitted capacity
20:40:59 <andythenorth> cb 32...32 day callback
20:41:18 <andythenorth> use cb 15 to set capacity to 'Dangerous'
20:41:29 <andythenorth> use cb 32 to give a random 1 in 50 chance of sinking
20:43:32 <fjb> Colliding ships would also be fun. :)
20:44:11 <Pikka> andy: "sink" the ship by setting the speed and running cost to 0, and changing the graphics? :P
20:44:14 <fjb> And a monster wave desaster.
20:45:38 <andythenorth> fjb: colliding ships would not be fun :(
20:45:46 <andythenorth> Pikka: it's an idea I guess
20:45:59 <andythenorth> annoying 'your vehicle is lost' messages though?
20:47:06 <Pikka> not if you have it turned off for stopped vehicles and manually stop the ship *shrug*
20:48:45 <andythenorth> OpenTTD is lacking in...evil
20:49:25 <andythenorth> although nfo is definitely not lacking in evil
20:50:36 <andythenorth> oops, forgot to go and code
20:52:13 <andythenorth> oh, something is much harder than I thought :O
20:54:39 *** boekabart has left #openttd
20:56:06 <Zuu> Personally I don't really mind. I know 29, 32, 33 very well. And if I am not mistaken the problem section is 31.
20:56:42 <Prof_Frink> Three Four Oh Seven.
20:56:45 <frosch123> Zuu: do you already know the new forums :p
20:56:46 <Zuu> Oh and 65 for NoAI of course.
20:57:28 <Zuu> frosch123: Which new forums? You mean the OpenTTD 4D forum?
20:57:46 <frosch123> no, grf releases and grf technical
20:58:09 <Zuu> Ah, those, nope. But I didn't really know the graphics forum number before anyways.
20:59:02 <Zuu> Though you was talking about hypotetically future forums :-D
20:59:30 <frosch123> the next one will be 69
21:00:04 <frosch123> that is easy, but the subject might be questionable
21:00:23 <andythenorth> I am trying to be unafraid of registers
21:00:27 <Zuu> Better take the easy questions first.
21:01:04 <andythenorth> oops, seen a mistake (embarassed)
21:01:09 <frosch123> andythenorth: don'T mix byte and word access
21:01:20 <andythenorth> that was one mistake
21:01:25 <andythenorth> there's another on the final line
21:02:01 <andythenorth> so chance 85 to 81 and remove
21:02:15 <andythenorth> so change 85 to 81 and remove some 00 00
21:02:30 <andythenorth> no reason to store a simple 0 / 1 flag in anything other than a byte?
21:02:50 <frosch123> well, it is always 1, so no reason to store it at all :p
21:03:14 <andythenorth> well there would be another varaction 2 setting 0
21:03:32 <andythenorth> it would be a branching chain.
21:03:50 <Eddi|zuHause> occasionally, booleans are stored in 32bit-words, for alignment purposes
21:03:55 <frosch123> there is also some comparison comparator which can be used for some simple stuff, but in most cases you are better off with a normal varact2 for the cases
21:05:15 <andythenorth> wiki tells me registers always have the size of 4 bytes. Do I need to care about that here?
21:05:46 <Eddi|zuHause> no, you can byte-access that dword, it will only fetch the lowest 8 bit then
21:06:23 <frosch123> store will always write all 4 bytes (zero expanded)
21:06:38 <frosch123> load will always read 4 bytes with the specified shift-right
21:07:22 <frosch123> (i.e. the 8 bit masking is done after the shifting)
21:08:10 <andythenorth> how do I read from a register? Sorry if I'm being dumb, but I can't see it in the wiki? I know how to do it for special cases like production callback, but they provide built in support, not roll your own...
21:09:24 *** nicfer1 has joined #openttd
21:09:53 <frosch123> note that storing allows indirect addressing, while loading does not :p
21:10:45 <andythenorth> what implications does that have when coding?
21:12:01 <frosch123> in most cases none, as indirect addressing is hardly needed
21:12:53 <andythenorth> is register 100 specifically for the text stack?
21:12:54 <frosch123> oh, and you could implement indirect addressing with some insane 7e procedure :roll:
21:13:09 <frosch123> 100+ registers are write-only
21:13:41 <frosch123> they are used for returning values from callbacks in addition to the usual 15bit result
21:14:03 <frosch123> some cb use them for the text stack, some cb use them for other stuff
21:14:15 <andythenorth> so registers <100 are all available? or are any reserved or strange?
21:15:06 <frosch123> registers 0x00 to 0xff are free for your use, but they are only hold their data during a single callback/spriteresolving
21:15:14 *** Polygon has joined #openttd
21:15:27 <frosch123> persistent storage only has 16 registers
21:15:34 <andythenorth> so to share something between callback I need persistent storage?
21:15:50 <andythenorth> is persistent storage shared across the whole grf?
21:15:59 <frosch123> no, 16bytes per industry
21:16:17 <andythenorth> well right now I only need 1 byte :)
21:16:17 <frosch123> err, 16 registers = 64 bytes
21:16:27 <frosch123> you can access them from both the industry and the tiles
21:17:10 <andythenorth> I just want to set a flag for 'cargo processed since cb 3D last called'
21:17:54 <TrueBrain> [21:36] <Sacro> TrueBrain: could start specifying 2k/XP/Vista/7 on the download page now > Request Forward > Rubidium
21:18:14 <frosch123> i hope you do not want to return cb36 "accept" only once until some cargo is processed
21:18:46 <frosch123> that will very unlikely result in "wanted" behaviour :p
21:20:04 * Pikka thinks newais still have a long way to go :D
21:20:18 <andythenorth> frosch123: what might go wrong?
21:20:40 <Pikka> eg, "don't try to drive an 800 ton train with a 600hp loco straight up a mountain"
21:21:27 <frosch123> andythenorth: the callback is called for every cargo packet. which means you will end up with about 1 to 5 per train and 1 to 20 per aircraft
21:21:40 <frosch123> (note: per train, not per wagon)
21:22:22 <Pikka> you mean every time it tries to unload something?
21:22:46 <frosch123> a cargopacket is a certain amount of cargo from one source, that is unloaded from one wagon in one gradual loading step
21:23:13 <frosch123> so, in most cases the loadamount, but if the cargo is from multiple sources, it could also only be 1
21:23:23 <Pikka> I'm not sure what you're getting at. :P still, try it and see, andy? :)
21:24:10 <Eddi|zuHause> unless you have extensive use of transfers or cargo destinations, most wagons will have 1 cargopacket, that is split up at every unloading step
21:25:47 <andythenorth> ach I've led you guys down the wrong path...it's cb 35 (production change) that I want the flag for :)
21:26:17 <andythenorth> "if cargo was processed since cb 35 last called, try production change, else ignore"
21:27:11 <frosch123> then you could as well use the variable for cargo_produced_last_month
21:27:19 <frosch123> or _this_month? no idea
21:27:26 <Pikka> except processed != produced :P
21:27:53 <andythenorth> Pikka: nail, meet head
21:28:05 <andythenorth> and one hit the other
21:28:21 <frosch123> well, i assumed everything that goes in, comes out somewhere
21:28:38 <andythenorth> nope, some just goes in. and stays
21:29:01 <Pikka> there is a "Date when cargo was last accepted". but it's short-format days-since-1920.
21:29:58 <andythenorth> well thanks chaps, but problem for tomorrow I think
21:30:15 <andythenorth> the code I have now works, but improving it would be...fairer to the player
21:34:36 *** TrainzStoffe has joined #openttd
21:41:53 *** TrainzStoffe is now known as Stoffe
22:01:17 *** Coco-Banana-Man has quit IRC
22:02:13 *** Brianetta has joined #openttd
22:17:38 *** SirSquidness has joined #openttd
22:59:49 *** lobster has joined #openttd
23:06:50 *** lobstar has joined #openttd
23:17:49 *** lobstar has joined #openttd
23:27:58 *** Rubix`` has joined #openttd
23:29:12 *** killahkrew has joined #openttd
23:31:14 *** Dred_furst has joined #openttd
23:33:19 *** Eddi|zuHause has joined #openttd
continue to next day ⏵
