IRC logs for #openttd on OFTC at 2018-06-21
⏴ go to previous day
00:30:48 *** techmagus has joined #openttd
01:50:53 *** FLHerne has joined #openttd
02:28:18 *** Thedarkb-X40 has joined #openttd
02:50:25 *** tokai|noir has joined #openttd
02:50:25 *** ChanServ sets mode: +v tokai|noir
03:52:01 *** supermop has joined #openttd
04:36:43 *** muffindrake has joined #openttd
05:02:16 *** ToffeeYogurtPots has quit IRC
05:02:34 *** ToffeeYogurtPots has joined #openttd
05:34:23 *** snail_UES_ has joined #openttd
06:26:05 *** Coobies has joined #openttd
06:27:45 *** innocenat_ has joined #openttd
06:29:35 *** Smedles_ has joined #openttd
06:30:39 *** mikegrb_ has joined #openttd
06:30:40 *** Laedek_ has joined #openttd
06:31:08 *** colde_ is now known as colde
06:31:08 *** innocenat_ is now known as innocenat
06:31:40 *** Wacko1976 has joined #openttd
06:31:52 *** innocenat is now known as Guest5956
06:32:47 *** techmagus has joined #openttd
06:33:49 *** TinoDidriksen has joined #openttd
06:34:09 *** OsteHovel has joined #openttd
07:15:30 *** anikhtos has joined #openttd
08:15:41 *** andythenorth has joined #openttd
08:23:10 *** ccfreak2k has joined #openttd
08:34:59 *** triolus has joined #openttd
08:42:48 <peter1138> So does nobody use the nightlies?
08:48:53 <andythenorth> there aren't any :P
09:05:09 <LordAro> someone should fix that
09:07:47 <peter1138> Yes, but as far as I know, nobody has complained.
09:08:12 <peter1138> Mind you, they might if NRT was merged ;)
09:09:26 <andythenorth> there is that yes :|
09:10:31 <peter1138> The PR needs to be updated.
09:11:53 *** gelignite has joined #openttd
09:14:20 <andythenorth> well my Russian is not great :P
09:17:15 *** WWacko1976-work has joined #openttd
09:25:38 *** chomwitt has joined #openttd
09:57:02 *** andythenorth has joined #openttd
10:19:18 <peter1138> andythenorth, does it matter?
10:19:33 <andythenorth> haven't looked yet
10:19:52 <andythenorth> if it's string conflicts, eh
10:19:53 <peter1138> All you do is apply the changes to your current version.
10:19:57 <Rubidium> ofcourse it matters, otherwise most of us wouldn't have the job we have now
10:20:39 <andythenorth> I'll rebase my fork later and see what happens
10:21:08 <andythenorth> also Bitbucket is returning ascii art on successful git push
10:21:10 <andythenorth> which is confusing
10:25:53 <peter1138> Well, the russian conflict is very simple and doesn't require understanding russian.
10:28:41 <peter1138> And the other one is pretty simple.
10:29:43 <peter1138> Hmm, can use the web editor? Hmm!
10:30:50 <peter1138> No you can't, because we use tabs as tabs, and the editor uses spaces as tabs :S
10:32:20 <peter1138> Actually you can, by copying and pasting existing tabs o_O
10:32:33 <andythenorth> that's how I do it in my editor :P
10:32:35 <peter1138> You can test the changes to ensure there's not a typo.
10:32:44 <peter1138> You *CAN'T* test the changes to ensure there's not a typo.
10:33:10 <andythenorth> what changed in your NRT branch? :o
10:33:15 <andythenorth> you're upstream to me :P
10:35:26 <andythenorth> maybe I should not touch that
11:10:06 *** ToBeFree has joined #openttd
11:10:18 *** supermop has joined #openttd
11:24:33 *** Thedarkb has joined #openttd
11:35:51 *** andythenorth has joined #openttd
11:38:27 *** gelignite2nd has joined #openttd
12:18:31 *** ToBeFree is now known as Guest5978
12:18:31 *** ToBeFree has joined #openttd
13:46:56 *** Flygon_ has joined #openttd
14:00:34 *** Stimrol has joined #openttd
14:02:44 *** supermop has joined #openttd
14:56:02 *** snail_UES_ has joined #openttd
15:54:31 *** Alberth has joined #openttd
15:54:31 *** ChanServ sets mode: +o Alberth
16:15:48 <andythenorth> it's like the past
16:15:54 <andythenorth> is belugas here too?
16:16:14 <DorpsGek> LordAro: Bjarni was last seen in #openttd 6 years, 37 weeks, 0 days, 13 hours, 57 minutes, and 7 seconds ago: <Bjarni> heh
16:16:25 <DorpsGek> andythenorth: darkvater was last seen in #openttd 12 weeks, 5 days, 1 hour, 7 minutes, and 35 seconds ago: <Darkvater> I'm looking for a distributed fs to setup my own private cloud using sbc's
16:16:39 <Eddi|zuHause> the past was never actually like the past
16:17:56 <nielsm> "missing person: left 12 weeks ago to search for a distributed file system"
16:19:23 *** ChanServ sets mode: +v tokai
16:30:33 *** ChanServ sets mode: +v tokai
16:34:09 <nielsm> I'm tempted to add a "bikeshed" tag to the "Update of README file" PR
16:34:39 <andythenorth> nah, it's not bikeshedding to have good docs :)
16:34:48 <andythenorth> although 'macOS' makes me sick in my mouth
16:34:57 <nielsm> whether to write macOS or Mac OS X is
16:35:08 <dihedral> darkvater - recently saw him on linked in
16:35:30 <LordAro> but equally i have absolutely no use for it
16:35:32 <andythenorth> and I use the damn thing
16:35:38 <dihedral> but we have a rubidium and a TrueBrain - we need not cherrish those old times
16:35:55 <andythenorth> shall we just highlight people annoyingly dihedral?
16:36:09 <LordAro> i choose andythenorth
16:36:32 <dihedral> i do not know what you mean, andythenorth
16:39:37 *** HerzogDeXtEr has joined #openttd
16:45:52 <Rubidium> andythenorth: why would that make you sick in your mouth? It's just nostalgia ;)
16:46:38 <andythenorth> mostly because it's 'Mac'
16:46:53 <andythenorth> and the shape of 'macOS' is horribly ugly
16:49:48 *** techmagus has joined #openttd
16:50:46 <Rubidium> oh joy... our "friends" at apple call it macOS now (previously Mac OS X and OS X)
16:51:04 <Rubidium> so useful to rename your product to have the same name as in the past
16:53:39 <Eddi|zuHause> the name is dead, long live the neame
16:55:30 <nielsm> windows went from version 4 to 2000 to NaN to 7, 8, 10
16:55:39 <nielsm> doesn't have to make sense
16:56:19 <Eddi|zuHause> yeah, but they didn't go back to Windows NT
16:56:27 <dihedral> servers still count 2008, 2012, 2016 ...
16:57:55 <Eddi|zuHause> what's worse is that there's an internal numbering and a marketing numbering that don't seem to have anything to do with each other, but both are used in different places to present to the user
16:58:16 <dihedral> devs and branding teams :-D
17:02:05 * dihedral started writing a tcp admin thingy in tcl...
17:02:15 <dihedral> just for the fun of it
17:06:02 *** Wormnest has joined #openttd
17:14:51 *** WWacko1976-work has quit IRC
17:23:10 *** Progman has joined #openttd
17:31:00 <nielsm> trying to add measurement of time taken to draw just viewports now...
17:36:59 *** ANIKHTOS has joined #openttd
17:39:39 *** HerzogDeXtEr has joined #openttd
17:42:29 *** TheMask96 has joined #openttd
17:43:37 <ANIKHTOS> nhello nielsm i watch the varibles as you suggeted
17:45:14 <ANIKHTOS> nielsm how must stress the statistic put on the cpu?? how much perfomance you loose when you run them ? is there a way to know??
17:46:15 <nielsm> drawing the windows that shows the framerate and performance numbers does cost CPU time, but taking the measurements doesn't take much in particular
17:52:47 <ANIKHTOS> so is there a way to calculate a ^ loss because of the measurements running??
17:53:18 <ANIKHTOS> it would give a better idea to slwo machines how the game runs for them
17:53:29 <andythenorth> one could write a tool to measure the performance...
17:55:31 <Eddi|zuHause> you can't know how much you change the value by measuring it. that's the core of Heisenbergs law
17:56:32 <ANIKHTOS> an estiamtion would nto be usefull ???
17:59:13 <andythenorth> run the same savegame on FFWD for timed runs
18:10:30 *** ANIKTHOS has joined #openttd
18:26:18 <ANIKTHOS> now the code is easier to read??
18:26:35 <ANIKTHOS> or still i am nto writitng the code as i ought to write it>>
18:32:35 <ANIKTHOS> in a game i either focus on cargo oron passangers
18:34:05 <nielsm> one I downloaded to use as an extreme stress test
18:38:54 *** Thedarkb has joined #openttd
18:39:00 <ANIKTHOS> so in the code i posted the bug is line 15 is not executed the first N times !?!??! but there is execution of code line 24 to 30!?!??!
18:39:51 <ANIKTHOS> now if i change the line 24 if (_dayh > _daym){ to if (_dayh == _daym){ line 15 is executed but now line 24 is not executed for N times
18:43:32 <nielsm> the ///< type of comments are normally only used for declarations, specifically fields in structs/classes and values in enums
18:43:41 <peter1138> Yeah but we're strange.
18:43:51 <nielsm> for regular code short comments just use //
18:44:04 <peter1138> I wonder if the measurements are starting to affect performance? :)
18:44:31 <ANIKTHOS> thanks nielsm i just added comments
18:44:37 <LordAro> hide the graphs and see what you get
18:44:47 <ANIKTHOS> i was waiting for the code to be finsihed to add them since i still play with it
18:44:49 <peter1138> LordAro, sure the graphs will
18:44:50 <nielsm> I hope the actual measuring doesn't have any appreciable effect, but painting the framerate GUI does take some time
18:44:56 <peter1138> LordAro, but I mean the actual measuring.
18:45:11 <peter1138> That level of detail does look really nice though :)
18:45:16 <LordAro> oh, well i'd hope not
18:45:35 <LordAro> what is it, an extra couple of stores every tick?
18:46:01 <nielsm> 64 bit integer math and some stores
18:46:02 <peter1138> Hopefully it's just one measurement per vehicle type? May not be though.
18:46:29 <nielsm> I don't know how fast/slow QPC is on windows but since it's for performance measurement, I assume it's not so slow as to affect performance significantly
18:46:52 <nielsm> nope, have to call it in Train::Tick, RoadVehicle::Tick, Ship::Tick, Aircraft::Tick
18:47:08 <nielsm> since they're all assembled in one large vehicle vector
18:47:41 <peter1138> That's a lot of measurements then.
18:48:02 <peter1138> I'll see in a second :)
18:48:12 <peter1138> Er, if this is pushed.
18:48:17 <nielsm> still each measurement should be trivial compared to the rest of the tick processing
18:48:40 <nielsm> ANIKTHOS, are you sure your variables are being initialized correctly?
18:49:43 <ANIKTHOS> well they are intiialized by copu earlier
18:50:39 <peter1138> Hmm, still hovering between 7.5 and 8fps on that same save.
18:50:40 <nielsm> oh oops, I mistaked a make
18:50:48 <peter1138> Maybe slightly slower but not massively.
18:51:36 <nielsm> if you collapse the main framerate window it will show just the basic rate in the caption, in debug mode I get a significant difference between full details and just basic rate rendering :P
18:52:55 <Eddi|zuHause> maybe the string assembly takes some time as well as the actual redrawing?
18:53:07 <peter1138> Hmm, gameloop is still 16ms with all vehicles paused
18:53:11 *** gelignite has joined #openttd
18:53:18 <peter1138> and train ticks is still 6ms with all trains paused o_O
18:53:46 <peter1138> I suppose there *are* a lot more rail vehicles.
18:59:47 <nielsm> there's more roadvehs than trains, but more train cars than roadvehs
19:00:09 <peter1138> Yeah, I said rail vehicles rather than trains.
19:00:14 *** frosch123 has joined #openttd
19:00:22 <peter1138> I typed "fps" in the console on a dedicated server.
19:00:40 <peter1138> Fails when no readings have been taken.
19:00:55 <Eddi|zuHause> why is a jurassic park tycoon-ish game age 16+?
19:01:02 <peter1138> and also there'll be no sound.
19:01:36 <Eddi|zuHause> why would a 12-year-old not enjoy a trex mowing through your visitors?
19:05:37 <nielsm> that should also let it skip video and sound when it's not doing that
19:06:15 *** Wacko1976 has joined #openttd
19:07:16 <peter1138> Hmm, network mode is totally fucked up with that savegame.
19:12:21 <peter1138> Maybe there's a problem with the network protocol when the server & client can't keep up.
19:18:17 <peter1138> nielsm, yeah that stops it crashing
19:18:33 <nielsm> pushed that fix for the fps command, also fixed it to not print rate numbers that are meaningless
19:18:57 <peter1138> Lots of fps values that aren't in the gui :p
19:20:24 <nielsm> I think I'll also add back a semi-collapsed state to the framerate window where it doesn't show all the data elements
19:25:33 <peter1138> You calculate the time of GameLoop()
19:25:46 <peter1138> However the actually gameloop is StateGameLoop()
19:26:46 *** sla_ro|master has joined #openttd
19:27:36 <peter1138> I guess the stuff in GameLoop() itself doesn't need to be measured, cos it's not the actual game loop.
19:28:42 <peter1138> And the network lag issue appears to be because the client is ALWAYS playing catch up with the server, and so barely has a chance to render.
19:30:29 <peter1138> There we go, moved the timer to StateGameLoop()
19:30:41 <peter1138> Simulation rate: 7.93 frames/s
19:30:50 <peter1138> Graphics frame rate: 0.60 frames/s
19:31:17 <peter1138> It was running at ~ 4 fps for a bit but then collapsed.
19:34:54 <LordAro> ooi, why "frames/s" rather than "fps" ?
19:36:02 <nielsm> because that's more stylish
19:50:07 *** andythenorth has joined #openttd
19:52:04 *** Thedarkb has joined #openttd
20:00:27 *** Supercheese has joined #openttd
20:02:20 <andythenorth> back to CSS for me then
20:20:01 <Eddi|zuHause> why does the scale say "s" when the label says "ms"?
20:27:42 *** supermop has joined #openttd
20:33:53 <nielsm> the peak label is always in milliseconds, the scale uses seconds when appropriate
20:34:00 *** planetmaker_ has joined #openttd
20:34:02 *** Wacko1976 has joined #openttd
20:36:12 <nielsm> on the idea that you want to see the full precision of the peak, but don't need to read superfluous zeroes on the axis
21:20:27 *** Supercheese has joined #openttd
21:23:07 <ANIKTHOS> does anyone want to re produce the bug and tell me if they also have it??
21:29:02 <peter1138> I want to eat these home-grown-from-my-garden strawberries, raspberries and gooseberries.
21:30:43 <nielsm> ugh I can't figure out how to handle "paused" states in a reasonable way...
21:31:01 <peter1138> If pause, don't update the measurements?
21:31:56 <nielsm> the gap in time causes things to be weird when you unpause
21:33:39 <andythenorth> peter1138: you can
21:34:11 <peter1138> andythenorth, I did!
21:42:06 <andythenorth> well I should fix NRT eh
21:42:16 <andythenorth> but it's recycling day :|
21:46:23 <Wolf01> <peter1138> I want to eat these home-grown-from-my-garden strawberries, raspberries and gooseberries. <- the whatberries?
21:54:18 <peter1138> Nah, no whatberries.
21:56:20 <Eddi|zuHause> nielsm: i think if paused the measurements should just be 0?
21:57:02 <Eddi|zuHause> the parts that aren't executed i mean. there's still like drawing and stuff going on
22:10:57 <nielsm> IMO it's the most appropriate way to handle it, since the game loop is supposed to be "in stasis" with no time passing in it
22:11:53 <nielsm> (also it took a bunch of care to get working so you better damn appreciate it!!! ;)
22:14:25 <Eddi|zuHause> if i'm watching my realtime applications, i want all the graphs be in sync
22:15:01 <ANIKTHOS> if ( _settings_game.game_creation.slow_time_factor==_dayn){ this line and this line if (_dayn == _settings_game.game_creation.slow_time_factor){
22:15:08 <ANIKTHOS> whats the difference??
22:16:26 <Eddi|zuHause> ANIKTHOS: make a printf() before this line, watch the values of the variables in an actual run
22:17:45 <ANIKTHOS> i change the status bar in the game and i display the variables i have created
22:18:18 <ANIKTHOS> but they ought to be the same??
22:19:18 <ANIKTHOS> how to i use th eprintif???
22:19:51 <nielsm> printf("dayn = %d\n", _dayn);
22:20:18 <nielsm> the \n at the end is to make a newline (line break)
22:20:45 <nielsm> the %d is a formatting code for "decimal" (i.e. integer value to be shown in normal base 10)
22:21:30 <nielsm> printf("stf = %d, dayn = %d\n", _settings_game.game_creation.slow_time_factor, _dayn);
22:21:56 <ANIKTHOS> where it will print them write them so i can see them??
22:22:31 <nielsm> I guess you only see that one if you're making debug builds, I don't know if you are or not
22:22:46 *** sim-al2 has joined #openttd
22:23:34 <ANIKTHOS> i shoudl type make -g??
22:24:41 <nielsm> I think it's a configure option to get debug builds
22:25:43 <ANIKTHOS> but i can already watch the variables what value they get
22:27:21 <ANIKTHOS> the bug disapeared when i reverse the logic question
22:29:13 <ANIKTHOS> for one reason it takes the value from this logic quesiton to make the bug apear that many times
22:29:37 <ANIKTHOS> reversing the question secodn part now is 0 so the bug will apear 0 times thus never
22:30:10 <LordAro> pretty sure make -g isn't a thing
22:30:42 <ANIKTHOS> i was readign gdt manual and it was make - somethign
22:30:58 <ANIKTHOS> make -o for optimized code
22:31:11 <nielsm> no those are for gcc and g++
22:31:34 <LordAro> you do not get to touch those directly
22:31:39 <LordAro> (yes yes i know about CFLAGS)
22:31:47 <LordAro> you should not need to
22:32:09 <peter1138> nielsm, assertion failed at line 656 of framerate_gui.cpp
22:32:13 <peter1138> newpoint.x <= lastpoint.x
22:32:39 <nielsm> ugh I thought I weeded out all possibilities of that
22:34:24 <nielsm> do you know which graph triggered it, and what situation?
22:36:16 <ANIKTHOS> well good news bug is suppressed to not apear, bad news i do not have any clue why it happened i will take the small win
22:44:10 <Eddi|zuHause> ANIKTHOS: try "./configure --enable-debug=1"
22:44:48 <Eddi|zuHause> also, the printf might be visible on the ingame console, not sure
22:45:17 <Eddi|zuHause> open the ingame console with the key left of [1]
22:45:44 <Eddi|zuHause> (actual key may differ between keyboard layouts)
22:46:26 <ANIKTHOS> si there a way to see if a lien is executed or not??
22:46:52 <Eddi|zuHause> that's what the printf is for
22:47:04 <Eddi|zuHause> or, i guess, you could learn gdb and set a breakpoint
22:48:07 <ANIKTHOS> lol i have not leart github
22:48:15 <ANIKTHOS> not leanred to make the diff files yet
22:48:24 <ANIKTHOS> and the list to learn get bigger and bigger
22:51:15 <ANIKTHOS> teh thigns shown in console can be saves in a log file??
23:33:24 *** Supercheese has joined #openttd
23:37:11 *** andythenorth has joined #openttd
continue to next day ⏵
