IRC logs for #openttd on OFTC at 2018-05-15
⏴ go to previous day
00:02:14 <peter1138> Can't be bothered putting the G27 on my desk. First-world problems.
00:13:51 <FLHerne> peter1138: If you want to rejig the map array, why not grab some of cirdan's work? He already did that :P
00:14:13 <FLHerne> (custom bridgeheads ftw)
00:16:20 <peter1138> Because it's very difficult to take *just* the map changes and nothing else.
00:17:40 <FLHerne> Ah, too much stuff on top of them?
00:18:37 <FLHerne> At least he seems to have a sane commit history, unlike a lot of the large patchsets
00:20:36 <FLHerne> I guess they depend on some of his other misc. cleanup changes, but aren't those worth having in their own right?
00:20:51 <peter1138> Just the map changes are massive, tbh.
00:22:46 <peter1138> The map accessors are now indirect to tile accessors.
00:27:00 *** chomwitt has joined #openttd
00:29:58 <peter1138> And anyway, the array isn't extended.
01:00:48 <peter1138> Just another Jay Foreman.
01:02:01 <peter1138> Seen him perform loads of times. Not seen his brother yet though.
01:31:04 *** supermop has joined #openttd
02:34:01 *** snail_UES_ has joined #openttd
02:48:39 *** Eddi|zuHause2 has joined #openttd
03:15:42 *** sim-al2 is now known as Guest2711
03:15:43 *** sim-al2 has joined #openttd
03:34:34 *** Prof_Frink has joined #openttd
04:32:13 *** muffindrake3 has joined #openttd
06:15:44 *** supermop has joined #openttd
06:55:23 *** Wacko1976 has joined #openttd
07:19:59 *** andythenorth has joined #openttd
08:14:42 *** synchris has joined #openttd
08:30:17 *** ToBeFree has joined #openttd
09:38:02 *** ChanServ sets mode: +v tokai
09:54:57 *** smoke_fumus has joined #openttd
10:21:09 <peter1138> I implemented HasBit() etc for std::bitset.
10:22:35 * peter1138 gets rid of snap. It basically installs a partitioned ubuntu in /snap.
10:26:01 *** andythenorth has joined #openttd
10:36:39 <peter1138> Hmm, my jenkins is not responding :S
11:01:40 <peter1138> LordAro, hmm, so your PR fails ;S
11:04:58 <peter1138> Doesn't seem to show an error message though.
11:10:12 *** Eddi|zuHause2 is now known as Eddi|zuHause
11:10:23 <Eddi|zuHause> <peter1138> PENISton i'm sure nobody ever made this joke before
11:10:45 <andythenorth> it's 25 years since I first heard it ;)
11:11:50 <peter1138> I have the best jokes.
11:12:02 <DorpsGek> andythenorth: pikka was last seen in #openttd 2 days, 0 hours, 55 minutes, and 11 seconds ago: <Pikka> not much sup... heading out in a minute
11:12:05 <peter1138> Also, make regression shows that I broke AIs.
11:13:47 * andythenorth ponders 'fund random industry' button
11:13:52 <andythenorth> for when there's too much money
11:14:03 <andythenorth> or I could just click the fund menu with my eyes shut :P
11:23:48 *** sim-al2 is now known as Guest2743
11:23:49 *** sim-al2 has joined #openttd
11:48:18 *** Laedek_ has joined #openttd
11:50:24 <LordAro> peter1138: yeah, i'm not entirely sure what i broke
11:55:20 <peter1138> Dunno, but it highlighted that I hadn't catered for the AI when changing command bit packing.
11:55:58 <peter1138> Maybe we should have accessors for DoCommand instead of calling it directly.
11:56:23 <peter1138> I should finish other stuff first :p
11:56:47 *** Lejving has joined #openttd
11:57:09 <Eddi|zuHause> moar indirection!
12:04:27 <peter1138> Ish. Compiler would most likely optimise it away.
12:48:41 <andythenorth> do I need giant boats?
13:01:33 <andythenorth> currently they top out at 720t
13:01:58 <andythenorth> FIRS has industries at 1500t / month
13:02:35 <Eddi|zuHause> that's still only 2 ships per month
13:03:10 <Eddi|zuHause> and there's no space restriction on ships
13:03:40 *** chomwitt has joined #openttd
13:08:19 <andythenorth> anything bigger doesn't fit in canals
13:08:25 <andythenorth> or throws out the relative scale
13:10:03 <Eddi|zuHause> <peter1138> Yet :) <-- as we discuss that for about 10 years already...
13:12:29 <peter1138> It's a very long yet.
13:13:37 <andythenorth> I learnt to think about 5 years as 'not very long' :P
13:14:14 <andythenorth> min. time to get anything done in software is about 1 year
13:17:15 *** JacobD88 has joined #openttd
13:25:02 <Eddi|zuHause> well, anything that is not done in 10 minutes
13:25:17 <Eddi|zuHause> (+ 30 minutes waiting for the computer to be ready to take your input)
14:04:37 *** Wacko1976 has joined #openttd
14:15:22 *** sla_ro|master has joined #openttd
14:18:22 *** supermop has joined #openttd
14:23:51 *** sim-al2 is now known as Guest2757
14:23:52 *** sim-al2 has joined #openttd
14:38:26 *** snail_UES_ has joined #openttd
14:43:01 *** supermop has joined #openttd
15:02:19 <peter1138> LordAro, how did you test PR 6781? OpenTTD segfaults on start for me.
15:02:44 <LordAro> pretty sure it ran for me
15:03:44 <peter1138> It segfaults before it gets to main()
15:04:55 <peter1138> strlen is wrong, because those are all compiletime, not runtime.
15:19:59 <LordAro> except the ones that are not
15:20:15 <LordAro> which is what the warning was about (mostly baseset::ini_set
15:47:14 <peter1138> LordAro, our saveload code checks that lengthof is the size of a pointer.
15:47:33 <peter1138> lengthof is defined, it's just not the runtime string length.
15:48:40 <LordAro> so OTTD performs the check the compiler is warning about and acts accordingly?
15:50:51 <peter1138> I don't have that version to test :p
15:51:03 <peter1138> I don't know what your compiler is warning about.
15:51:09 <peter1138> We define lengthof ourselves in stdafx
15:51:20 <peter1138> Maybe it's being "helpful" and overriding it?
15:51:21 <LordAro> lemme find the output
15:51:37 <peter1138> sizeof(x) / sizeof(x[0])
15:52:00 <LordAro> yeah, it's warning about dividing by sizeof of pointer type
15:53:14 *** Wacko1976 has joined #openttd
15:57:14 <peter1138> The correct solution may just be sizeof() instead of lengthof()
15:57:46 <peter1138> sizeof/lengthof is the same for strings, because each element is only a byte anyway.
15:59:18 <LordAro> unless it's a pointer type instead of an array type
15:59:20 <LordAro> which is the issue here
16:01:28 <peter1138> Try with sizeof instead of lengthof in the same places.
16:02:20 <peter1138> But yeah, this is not runtime, hence strlen will not work.
16:02:45 <SpComb> this sounds like a fun discussion
16:03:15 <peter1138> And wouldn't work on loading, because the strlen of the current value is not relevant.
16:03:47 <LordAro> but i'm not convinced sizeof/lengthof is the correct solution either
16:04:34 <peter1138> Well, this is something that has worked since forever.
16:05:40 <LordAro> i'd want to look at this "saveload code checks that lengthof is the size of a pointer" bit
16:06:46 <LordAro> i'll look later this evening
16:06:50 <peter1138> return sld->size == sizeof(void *) || sld->size == sld->length;
16:07:14 *** supermop has joined #openttd
16:10:48 <peter1138> Also, SLE_VAR_STR is for pointers
16:11:01 <peter1138> SLE_VAR_STRB is for arrays
16:11:51 <peter1138> So basically the code knows. lengthof/sizeof are interchangeable here except for your new warning.
16:22:33 <LordAro> need to reevaluate anyway
16:24:43 *** davidofmk771 has joined #openttd
16:56:36 *** andythenorth has joined #openttd
17:09:13 *** Wormnest has joined #openttd
17:11:07 *** Wacko1976 has joined #openttd
17:12:38 *** Thedarkb-X40 has joined #openttd
17:54:49 *** Wacko1976 has joined #openttd
18:12:31 *** TheMask96 has joined #openttd
18:40:20 *** Thedarkb-X40 has joined #openttd
19:12:25 *** HerzogDeXtEr has joined #openttd
19:18:51 *** andythenorth has joined #openttd
19:23:16 *** rocky1138 has joined #openttd
19:45:31 *** LANJesus has joined #openttd
19:51:44 *** Progman has joined #openttd
20:04:36 *** frosch123 has joined #openttd
20:20:25 *** Lejving has joined #openttd
20:24:29 <andythenorth> Iron Horse is a lot of trains to draw :P
20:25:47 <frosch123> how many people would notice when the graphics are the same once the train leaves the depot?
20:36:37 *** gelignite has joined #openttd
20:37:13 <V453000> proper grafix idea frosch123
20:57:18 <Eddi|zuHause> frosch123: a value less than "all" but more than "none"
21:09:04 *** Wacko1976 has joined #openttd
21:13:22 <andythenorth> frosch123: in the / \ views, people wouldn't notice
21:13:32 <andythenorth> and definitely not the | view
21:13:52 <andythenorth> horse is deceptively playable, because I'm drawing it over Horse 1 sprites :P
21:14:08 <frosch123> you already have the --- sprites
21:14:08 <andythenorth> if I fixed all the broken offsets, nobody might ever notice :P
21:14:20 <andythenorth> there was some grf with alignment vehicles?
21:14:29 <frosch123> if you wait for people to notice you also get feedback which engines are used :p
21:14:43 <andythenorth> is there any chance we could (a) find out why the sprite aligner is broken and (b) fix it?
21:14:46 <frosch123> there is debugveh on bananas
21:15:02 <andythenorth> the numbers sprite aligner displays have no obvious connection to those in the grf
21:15:17 <frosch123> remove the "-c" option from nml
21:15:19 <Eddi|zuHause> that's how modern game development works? you work around the biggest bugs and then ship it, hoping nobody notices the smaller ones?
21:15:39 <frosch123> no cropping, bigger grf, same offsets
21:15:54 <andythenorth> frosch123: can I do that permanently? Or do I have to piss around with a makefile arg? :)
21:16:37 <Eddi|zuHause> andythenorth: it's not meant for "permanent", just debugging purposes
21:16:55 <andythenorth> so I have to manually edit the makefile every time? o_O
21:17:08 <Eddi|zuHause> andythenorth: and you can just "alias nmlc=nmlc -c" or something
21:17:29 <Eddi|zuHause> possibly with some ''
21:18:53 <frosch123> Eddi|zuHause: how would that work?
21:19:42 <andythenorth> -c takes 6.1MB down to 5.2MB
21:19:44 <frosch123> andythenorth: don't you already have a variable for the fork bomb?
21:19:46 <andythenorth> so not critical, but worth it
21:19:53 <andythenorth> I have a var for the fork bomb yes
21:19:57 <Eddi|zuHause> another option is if the makefile is prepared with an "NMLCARGS" variable, you can do "NMLCARGS=-c make"
21:20:10 <frosch123> Eddi|zuHause: that is the wrong way around
21:20:25 <andythenorth> or I just do this once, fix the offsets, restore -c I guess
21:20:25 <Eddi|zuHause> both should work?
21:20:26 <frosch123> you want a variable to disable "-c", not to enable it
21:20:55 <andythenorth> or I have 'make bundle_tar' use -c
21:21:01 <andythenorth> and plain 'make' doesn't
21:22:15 <Eddi|zuHause> frosch123: maybe there should be a "--no-crop" option?
21:22:26 <andythenorth> I thought there was? o_O
21:22:33 <Eddi|zuHause> that overrides a "-c" option
21:29:34 <andythenorth> so I removed -c but the offsets are still odd
21:29:50 <frosch123> are they multiplied by 4?
21:30:10 <frosch123> i don't remember whether we display 1x or 4x offsets or both or depending on zoom level
21:30:23 <frosch123> maybe it's also the gui zoom setting
21:34:13 <andythenorth> you concluded same, they're multiplied by 4
21:34:24 * andythenorth was busy making screenshots :P
21:34:44 <andythenorth> UI zoom level doesn't seem to affect them
21:35:54 <frosch123> ok, so we only show 4x. i guess you closed some feature requests about showing 1x offsets :p
21:36:44 <andythenorth> fix, or I just divide by 4?
21:40:57 <peter1138> Is the sprite aligner wrong then?
21:41:41 <frosch123> peter1138: it's fine if you do 4x sprites :)
21:41:47 <frosch123> but 1x people are confused
21:41:48 <peter1138> As everyone should!
21:42:13 <peter1138> Judging by that bounding box, those offsets are wrong anyway.
21:43:28 *** muffindrake has joined #openttd
21:45:27 <andythenorth> they are wrong :P
21:45:43 <andythenorth> if they were right...I would be talking about something interesting
21:45:52 <andythenorth> like how many railtypes we can have
21:53:43 <frosch123> there should be an option to pick which 3 railtypes to use: if you want elrail you can choose to merge monorail into maglev or maglev into monorail
21:54:54 <frosch123> if i would post that on the forums, noone would understand it :)
21:55:41 <andythenorth> doesn't 3 waste half a bit? o_O
21:55:53 <andythenorth> we could use that for something else maybe? o_O
21:58:39 <andythenorth> 10 years of playing ottd is not enough
21:58:53 <andythenorth> today I found you can ctrl-click on the station name a vehicle is heading for
21:59:08 <andythenorth> (in the main vehicle window)
22:01:02 <andythenorth> boringly, I know most of them :P
22:01:06 <andythenorth> "Show vehicle load and capacity of multiple vehicles" is new to me though
22:01:36 <peter1138> Yeah we never copied that unified maglev feature from ttdpatch...
22:01:40 <frosch123> it's the unique ctrl+right click, i take full responsibility for that :)
22:04:38 <andythenorth> that's just to troll mac users
22:04:45 <andythenorth> ctrl-click is a right click on mac os
22:06:50 <frosch123> also notice that "ctrl click on engine in depot" says "reverse engine direction", not "switch livery" :p
22:07:10 <peter1138> That'll be a bug report coming in soon then.
22:09:47 <andythenorth> so the patch to incremement a counter (0-15) when ctrl-clicking...
22:09:58 <peter1138> 16 reversal states?
22:10:15 <andythenorth> A-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B
22:10:45 <andythenorth> probably needs shift-ctrl-click to decrement? o_O
22:10:59 <frosch123> can you also long-click to morse code a specific livery number?
22:12:02 <frosch123> andythenorth: how about a train configuration gui similar to grf parameter gui? :p
22:12:14 <andythenorth> I like mystery clicking :P
22:12:18 <V453000> It's so good to not be the only one trying to fuck over the specs to get some greater good (tm) functionality
22:12:24 <andythenorth> it's very direct, even though you don't know what you're getting
22:12:29 <andythenorth> ctrl-clicking is fun
22:12:38 <andythenorth> picking from dropdown menus is dull :P
22:17:19 <frosch123> what if the dropdown would show the sprite? and what if the dropdown was no vertical list, but a circle?
22:17:45 <andythenorth> does it include a unicorn :)
22:18:06 <andythenorth> oops I meant slug :P
22:18:56 <Eddi|zuHause> andythenorth: how about random liveries, but ctrl+click rerolls the random bits?
22:19:24 <andythenorth> that would be good trolling though
22:20:21 <Eddi|zuHause> well, that would simply be an UI change, instead of adding some mysterious counter thingie
22:20:41 <frosch123> the other day i wondered about improving map compression by removing all tile random bits, and instead store the number of rerandomisations
22:20:49 <Eddi|zuHause> which would involve changing vehicle storage code, GRF interface code, nfo specs and nml implementation
22:21:19 * peter1138 tests with 255 railtypes.
22:21:26 <andythenorth> imagine the abuses a counter could be put to
22:21:27 <peter1138> std::bitset testing :p
22:22:00 * andythenorth tries to remember how bitmasks work
22:22:14 <frosch123> grf-defineable toolbar ribbons for "formatting" train appearance
22:22:14 <Eddi|zuHause> andythenorth: rather add local storage registers for vehicles
22:22:38 <andythenorth> Eddi|zuHause: it's more appealing to do something that isn't proper
22:22:47 <andythenorth> it's a better meta-game
22:23:17 <frosch123> andythenorth: add livery wagons
22:23:33 <frosch123> their purpose is to change the livery of the other vehicles in the chain
22:23:34 <Eddi|zuHause> andythenorth: but you can then have your counter by adding a callback that is called on ctrl+click, which reads the register, adds one, and writes the register back
22:23:48 <peter1138> toolbar ribbons! but it doesn't need a spec cos modern language
22:23:54 <andythenorth> Eddi|zuHause: yes that would work
22:24:01 <andythenorth> it's just a bit too correct
22:24:17 <andythenorth> frosch123: the liveries I did already do that :P
22:24:19 <peter1138> I think 64k railtypes is not possible.
22:24:23 <frosch123> peter1138: did actually any software make good use of ribbons beyond office?
22:24:26 <andythenorth> such sad news peter1138
22:24:31 <andythenorth> the game will now die
22:24:37 <peter1138> Not without extending the network command interface.
22:24:44 <peter1138> frosch123, you suggest that was a good use of it?
22:25:01 <andythenorth> is the ribbbon the toolbar at the top of my Word window?
22:25:07 <frosch123> it added hierarchical hotkeys
22:25:13 <frosch123> kind of like vim :p
22:25:24 <Eddi|zuHause> i have no clue what ribbons are in this context
22:25:39 <andythenorth> so a ribbbon is a toolbar, but with MS marketing around it?
22:26:15 <frosch123> a ribbon is a tabbed toolbar that megers classic toolbars and dropdown menus
22:26:53 <peter1138> So std::bitset works.
22:27:22 <peter1138> I cheated and defined HasBit/SetBit/ClrBit for bitsets, instead of modifying all the code elsewhere to use 'native' calls.
22:27:28 <frosch123> are there std::bitsets in saveload code yet?
22:27:54 <peter1138> Mainly because the existing HasBit compiles fine for bitsets, though won't work properly beyond 32 bit.
22:28:04 <peter1138> railtypes is never stored.
22:28:31 <frosch123> peter1138: i read using "std::begin(container)" is more modern than "container.begin()"
22:28:42 <peter1138> Although there is a left over at the moment where it's loaded because it used to be stored. It should just be a CONDNULL.
22:28:52 <peter1138> Sadly, I can't just commit a quick fix now ;(
22:28:56 <frosch123> so is "HasBit(set)" more modern than "set.HasBit()"?
22:57:13 *** andythenorth has left #openttd
23:03:04 <Eddi|zuHause> if it were python, set.HasBit() would be the same as HasBit(set)
23:32:24 *** sim-al2 has joined #openttd
23:48:13 *** gas1[m] has joined #openttd
continue to next day ⏵
