Log for #openttd on 29th April 2018:
Times are UTC Toggle Colours
00:01:28  <peter1138> TrueBrain, hmm, git commit 60f0610a851 is strange, did moving that equation make a big CPU difference I wonder?
00:42:44  <peter1138> bc
00:53:00  <Eddi|zuHause> ENotAShell
00:56:25  *** Flygon has joined #openttd
01:01:12  *** Thedarkb1 has quit IRC
01:08:55  <peter1138> Well shit, how is that happen. I better go to bed.
01:21:43  *** Thedarkb1 has joined #openttd
01:29:42  <Thedarkb1-X40> Has anyone ever tried a hub and spoke style network?
01:34:10  *** iSoSyS has joined #openttd
01:35:42  *** ToBeFree has quit IRC
01:50:14  *** glx has quit IRC
02:03:18  <Eddi|zuHause> a spooky network?
02:03:28  <Eddi|zuHause> i don't think i had that
02:03:59  <Eddi|zuHause> also, i think it's probably a terrible idea
02:32:35  *** Coobies has joined #openttd
02:37:50  *** Cubey has quit IRC
02:47:51  *** iSoSyS has quit IRC
02:51:54  *** muffindrake3 has joined #openttd
02:53:47  *** muffindrake2 has quit IRC
03:14:25  *** HerzogDeXtEr1 has joined #openttd
03:20:07  *** HerzogDeXtEr has quit IRC
04:03:55  *** sim-al2 has joined #openttd
04:48:46  *** Coobies has quit IRC
05:08:58  *** snail_UES_ has quit IRC
05:31:07  *** Alberth has joined #openttd
05:31:07  *** ChanServ sets mode: +o Alberth
05:52:09  <Rubidium> woopsie... forgot that my MIPS QEMU was still in temp when I shut my computer down yesterday :(
05:53:46  <Eddi|zuHause> you... shut down your computer?
05:55:27  <Rubidium> yeah
05:58:03  <Alberth> :(
05:58:03  *** Wacko1976 has joined #openttd
06:03:04  <Rubidium> ah well, strgen resulted in the same files so it should (tm) work fine
06:10:40  *** nielsm has joined #openttd
06:19:00  *** andythenorth has joined #openttd
06:19:04  <andythenorth> o/
06:32:28  *** Progman has joined #openttd
06:39:50  *** synchris has joined #openttd
06:42:52  *** Wacko1976 has quit IRC
06:55:10  <andythenorth> @seen pikka
06:55:10  <DorpsGek> andythenorth: pikka was last seen in #openttd 18 hours, 53 minutes, and 7 seconds ago: <Pikka> at that point, why not just turn plane crashes off?
07:49:51  *** Pikka has joined #openttd
07:50:15  <andythenorth> see
07:57:46  <Pikka> I do see
07:58:30  <andythenorth> my train set doesn't work
07:58:47  <andythenorth> I played a game to try it
08:00:49  <Pikka> what's wrong with it?
08:01:17  <andythenorth> flippant answer: no super deltic (4,400hp proposed but never built)
08:01:20  <andythenorth> serious answer
08:01:54  <andythenorth> 'one obvious engine choice' works brilliantly, except when it gets monotonous
08:02:23  <andythenorth> for pax there are more options, electric, diesel, metro, railcar, etc
08:02:41  <andythenorth> but for freight, I do a lot of TL5, and it's all moguls or 37s
08:04:17  <Pikka> isn't that going to be the case even with "realistic" sets? I mean the more powerful engines are only use for higher TL + weight multiplier
08:04:28  <andythenorth> actually in my test game it's not all 37s, it just seems that way
08:04:32  <andythenorth> 13 out of 53 trains
08:04:53  <andythenorth> I think there's a difference between 'one obvious choice' and 'only one choice'
08:05:22  <andythenorth> hmm
08:05:40  <Pikka> I'm not sure it's possible to not have an "obvious choice" in TTD
08:05:54  <andythenorth> it's possible to obscure the choice :)
08:06:13  <Pikka> "realistic" sets have an advantage in that players will choose vehicles for reasons other than game mechanics
08:06:39  <andythenorth> see NARS 2 :)
08:06:50  <Pikka> yep
08:07:30  <Pikka> whereas something like Pineapple Trains, where the vehicles are completely fictitious... it already has a low number of locos, and adding more would be pointless because there's just not enough gameplay niches
08:07:44  <Pikka> they'd either not be used, or they'd cause something else to not be used
08:08:02  <andythenorth> I keep poking at adding a 1200hp engine, below the 37
08:08:10  <andythenorth> but it would be under-powered for TL5
08:08:19  <andythenorth> maybe I just need another livery :P
08:08:39  <andythenorth> I banned engine liveries
08:09:11  <Pikka> Rats... nice for shorter trains, and little passenger trains @ 90mph, but not enough power for heavy freight. They work. :)
08:09:33  <Pikka> or double headed on heavy freight, doubleheading is always nice
08:09:35  <andythenorth> kind of got the OP 73 for that
08:10:01  <andythenorth> maybe I add a joker steam engine around 1950 or so
08:11:31  <andythenorth> I added a joker heavy electric :)
08:11:48  <andythenorth> doesn't align with the standard intro dates or model life
08:13:36  <Pikka> "generations" can get a bit stale
08:13:47  <andythenorth> they do
08:13:52  <andythenorth> especially when it's just another 37
08:14:06  <Pikka> oh dear
08:14:11  <Pikka> civilai's gone rogue
08:14:15  <andythenorth> generations aren't a flawed concept
08:14:18  <Pikka> it's got six trains on a two-train line
08:14:26  <andythenorth> child A reports that Civil AI is a good competitor btw
08:14:35  <andythenorth> he says when it gets ships, it will be super hard and OP
08:14:39  <Pikka> :D
08:14:48  <Pikka> I'd better add ships then
08:14:58  <andythenorth> hmm, is a garratt needed?
08:15:03  <Pikka> adding more variety to the train networks atm
08:15:04  <andythenorth> is it ever needed? :P
08:15:12  <Pikka> I don't know
08:15:17  <Pikka> they're a bugger to draw though :P
08:15:26  <andythenorth> they are, I tried 2 for the african roster
08:15:39  <andythenorth> file under DVTs
08:15:40  <andythenorth> just no
08:16:16  <Pikka> unless you really abstracted the articulation, rather than trying to make them look realistic
08:16:57  <Pikka> still need a quintuplex for NA ;)
08:18:59  <andythenorth> I should make a nuts style set
08:19:19  <andythenorth> all fake
08:20:27  <Pikka> you shouldn't, tbh. It's fun to play with concepts but I think the "model train" aspect is what makes TTD work. People like trains they recognise.
08:20:40  <andythenorth> me too
08:20:48  <andythenorth> but I keep finding pages about 'never built' UK trains
08:21:10  *** FLHerne has joined #openttd
08:21:16  <andythenorth> the problem with 5000hp diesels in 1970 is where to go in 2020 with progression
08:21:34  <andythenorth> anyway, is it one of these as a joker?
08:22:02  <Pikka> definitely, Q1s are great
08:22:09  <Pikka> and an original TT vehicle :P
08:22:11  *** chomwitt has joined #openttd
08:23:51  * andythenorth adds it to 'todo'
08:25:31  *** debdog has quit IRC
08:30:11  <andythenorth> hmm maybe FLHerne's suggestion for mixed traffic, round 3
08:33:58  <andythenorth> or it's quite enough already
08:34:03  <andythenorth> 1989 ^
08:35:22  *** FLHerne has quit IRC
08:36:16  <Alberth> o/
08:37:59  <andythenorth> ho
08:39:19  *** Wolf01 has joined #openttd
08:39:55  <Wolf01> o/
08:40:44  <Wolf01> TrueBrain what are you doing with datacenters? :D "Repairs are ongoing after a massive power outage in the Amsterdam region that affected many services."
08:45:21  <Rubidium> sounds like a bad datacenter
08:47:47  <Rubidium> can't even spot a dip in Amsterdam's internet exchange throughput
08:48:40  <Wolf01> Telegram is down, they are working on it
08:48:47  <Eddi|zuHause> there was a power supply "hiccup" at DE-CIX a few weeks ago
08:49:44  <Eddi|zuHause> as in "the backup generators ran out of fuel after 2 hours"
08:50:02  <Alberth> Airport Schiphol had very long queues due to power outage
08:52:13  *** andythenorth has quit IRC
08:52:37  <Rubidium> why does checkin seems to be such an achilles heel for airports?
08:53:28  * Eddi|zuHause mumbles something about bombs and skyscrapers
08:54:46  <Wolf01> Now I'm doing something I didn't do for a LONG time, I'm reinstalling a WinXP laptop
08:55:24  <Rubidium> Eddi|zuHause: lately I'm only hearing of huge delays on airports because the checkin system did not function for one reason or another
08:55:27  *** Zexaron has joined #openttd
08:55:35  <Wolf01> Couldn't wait to get all those wild viruses again
08:56:53  <Eddi|zuHause> Rubidium: would you rather hear of delays because the air traffic control system didn't work?
08:58:43  <Rubidium> Eddi|zuHause: nope, but ATC not functioning has way more safe guards than airports closing entry roads due to check-in not working and due to that people walking with suitcases on the high way...
08:59:43  <Eddi|zuHause> anyway, i think the checkin system is probably uniquely qualified for huge delays because it a) is not safety relevant, so less rigorously tested and checked, and b) complex enough that _something_ will probably break (checking tickets against the database, routing baggage along conveyor belts, ...)
08:59:51  <Rubidium> those ATC people are trained to still function, although less efficiently, without radar and other electronic help. Only when the radio fails there's a huge problem, but even then planes can coordinate between eachother
09:01:41  <Rubidium> i.e. the pilots can and will communicate with eachother, but on the highway... they'll have barely any notice of people walking there
09:03:00  <Eddi|zuHause> humans do weird stuff in large groups
09:06:12  *** andythenorth has joined #openttd
09:10:54  *** FLHerne has joined #openttd
09:14:45  *** Gja has joined #openttd
09:16:25  *** andythenorth has quit IRC
09:23:37  *** frosch123 has joined #openttd
09:29:09  <Alberth> o/
09:29:19  <frosch123> moi
09:42:39  *** Wormnest has joined #openttd
09:53:33  *** andythenorth_ has joined #openttd
09:54:18  <andythenorth_> 90mph gronk?
09:55:08  *** andythen_ has joined #openttd
09:55:30  *** sla_ro|master has joined #openttd
09:57:58  *** andythen_ has quit IRC
10:02:22  *** andythenorth_ has quit IRC
10:14:31  <nielsm> hmm is there any patch that extends "disallow train reversing in stations" with a third option, where multiple units (with cabs at both ends) are allowed to reverse in stations but non-MUs are not?
10:16:12  *** chomwitt has quit IRC
10:17:14  <frosch123> there are newgrf which set the speed limit to something like 30 km/h when driving in a weird direction
10:19:13  <nielsm> that would have to be combined with something that reverses the entire consist when the train reverses at end of line, right?
10:19:52  <nielsm> meaning you'd get consists with no engine at the front
10:27:11  <frosch123> let's say, it is comilicated :)
10:27:14  <frosch123> +p
10:27:23  <TrueBrain> [02:01] <peter1138> TrueBrain, hmm, git commit 60f0610a851 is strange, did moving that equation make a big CPU difference I wonder? <- I think this came from a profile; so most likely it was. We did a few optimizations based on profiles. I rarely do any optimizations unless I know it matters :)
10:38:59  <TrueBrain> frosch123: OS check, do we want to keep support for the following: FreeBSD, DragonFly, OpenBSD, NetBSD, HP-UX, MorphOS, BeOS, Haiku, SunOS, Solaris, Cygwin, OS2, DOS, WINCE, PSP ?
10:41:11  <frosch123> in the past two years we had contributions for some BSD and OS/2
10:41:39  <frosch123> dos is only toyed on by rb and roboboy (or whatever nick he uses these days)
10:42:15  <frosch123> hpux, morphos, beos, haiku, sunos, solaris, wince and psp i have not seen in the past 10 years
10:42:32  <frosch123> i have no idea whether people use cygwin, or rather mingw
10:42:40  <TrueBrain> we can aks glx that
10:42:46  <TrueBrain> but I havent seen Cygwin in a long long time :)
10:43:18  <TrueBrain> okay, so I will see if I can make Dockers for the BSDs. OS/2 .. well, nice gimmick, find by me (does it have a c++11 compiler?)
10:43:25  <TrueBrain> DOS I will ask Rb, but I think that is fine too
10:43:38  <TrueBrain> means I am going to make a commit to remove the others; espcially those with a lot of special code
10:46:37  <Alberth> you can also drop anything without c++11 support, I guess
10:46:48  <TrueBrain> which is a bit hard to explore :)
10:46:51  <Alberth> not sure if that helps :)
10:47:03  <nielsm> just check if clang exists for the platform?
10:47:03  <TrueBrain> But stuff like PSP and WinCE are just experiments gone bad ...
10:47:10  <TrueBrain> I checked in PSP :P
10:47:19  <TrueBrain> nielsm: not a bad idea tbh
10:49:02  <nielsm> if you have clang, X11, and OSS, I'd wager most real unices can be supported without any special handling
10:49:25  <TrueBrain> and cmake :D
10:53:34  <frosch123> TrueBrain: i guess when it comes to gcc, we only support g++ >= 6.0
10:54:04  <frosch123> so, all the gcc 3/4 things can leave
10:54:25  <TrueBrain> that would be a nice cleanup too
10:57:33  <frosch123> though RHEL7 only has gcc 4.8
10:57:44  <frosch123> which supports c++11 inofficially
10:57:50  <frosch123> but no gc++14
10:58:31  <TrueBrain> but we were going to do c++11, not? or c++14?
10:59:01  <frosch123> i would prefer c++14, it finishes stuff that c++11 started
10:59:16  <frosch123> i would say no to c++11 closures :p
10:59:18  <TrueBrain> so we drop RHEL7? :)
10:59:29  <frosch123> yep :)
10:59:39  <TrueBrain> dont know what closures are :P
10:59:56  <frosch123> lambda functions or annoynmous functions
11:00:35  <frosch123> std::bind, but sane
11:03:31  <LordAro> you can get newer compilers for RHEL
11:08:52  *** Thedarkb1 has quit IRC
11:13:29  *** andythenorth has joined #openttd
11:19:27  *** Thedarkb1-X40 has quit IRC
11:25:10  <FLHerne> andythenorth: Current FISH 2 (2.0.3) on Bananas is named 'Squid Ate Fish r1687' in the selection menu
11:25:16  *** Zexaron has quit IRC
11:25:24  <FLHerne> This is a bit confusing
11:25:30  <andythenorth> yeah
11:25:39  <andythenorth> the names seem to be really unreliable
11:25:41  <andythenorth> tbh
11:25:54  * FLHerne just spent a few minutes trying to get rid of whatever weird devel version was masking the stable release, until realising...
11:26:38  <andythenorth> I've never worked out where OpenTTD gets the names from
11:26:56  <andythenorth> we rewrote a newgrf makefile to try and eliminate this issue
11:27:08  <andythenorth> but still happens, had a broken FIRS release recently due to this
11:28:18  <andythenorth> I won't fix Squid, it's legacy now
11:28:21  <andythenorth> just unfortunate :P
11:30:34  *** tokai|noir has joined #openttd
11:30:34  *** ChanServ sets mode: +v tokai|noir
11:32:41  <Wolf01> Ha firefox... started from FF15, updated to 43, to 47, to 52esr and now is dowloading another update...
11:32:58  <Wolf01> Downloading too
11:33:27  <andythenorth> FLHerne: if there was a similar-sized alternative to the Growler, what would it be?
11:33:37  <andythenorth> introduced sometime 1950-1975
11:34:20  <FLHerne> Hymek?
11:34:40  <FLHerne> 31s for ubiquity, but a bit underpowered
11:34:43  <FLHerne> (in RL, too...)
11:34:58  <andythenorth> needs to be 8/8 long, at least if you squint
11:35:44  <FLHerne> Would say 40, but a squashed 40 would be indistinguishable :P
11:36:04  <andythenorth> yeah I had same thought
11:36:04  <FLHerne> It's annoying that all the Modernisation Plan locos were so crap...
11:36:12  <andythenorth> well Horse alllows faking
11:36:24  <andythenorth> co-bo :P
11:36:31  <andythenorth> too-long class 33
11:37:27  *** tokai has quit IRC
11:37:34  <FLHerne> You don't actually have a 47 yet, do you?
11:37:39  <FLHerne> But you do have 57s, so meh
11:38:15  <FLHerne> Warships aren't really comparable
11:39:44  <andythenorth> the Dragon / Wizzo is kind of a warship
11:39:49  <andythenorth> also kind of a western
11:40:03  <andythenorth> could do a 50 and hobble the stats
11:40:17  <andythenorth> but it would just look like the thunderbird
11:40:37  *** gelignite has joined #openttd
11:40:40  <FLHerne> I guess technically, some of the Standard steam locos are 'sometime 1950-...'
11:40:46  <FLHerne> (but no)
11:41:06  <Pikka> GT3 o/
11:41:27  <andythenorth> class 38?
11:42:05  <andythenorth> Pikka: you actually doing GT3? o_O
11:42:17  <Pikka> no :P
11:42:32  <FLHerne> Deltic for the express, and cut back the Dragon to be mixed-traffickey?
11:42:51  <FLHerne> But then the Deltic still looks very 37ish at TT scale..
11:42:54  <andythenorth> it's like that rule in Bill and Ted
11:42:56  <andythenorth> "No Deltics"
11:43:16  <andythenorth> oh it was Wayne's World but eh
11:43:18  <FLHerne> UKRS2 has a GT3
11:43:24  <FLHerne> Not sure why...
11:43:29  <andythenorth>
11:43:39  <andythenorth> Led Zellepin policy
11:44:19  <FLHerne> (correction, it's in the extension)
11:44:52  <andythenorth> current lead contender is Class 75
11:44:57  <andythenorth> and introduce it in 1975
11:45:11  <andythenorth> dunno what it would look like though
11:45:12  <FLHerne> Those funny Ivatt prototypes would be about right
11:45:29  <FLHerne> (would still have the EE-family-look problem)
11:45:41  <andythenorth> there's the big Southern 1
11:46:00  <andythenorth>
11:46:06  *** Gja has quit IRC
11:46:14  <andythenorth> right power band, unusual shape
11:46:32  <andythenorth> but wrong era
11:47:01  <andythenorth> I'd rather do the bulleid 0-6-0 in 1945 or so, then do a replacement for it in 1975
11:47:22  <FLHerne>
11:48:59  <Pikka> that'a a nice prototype Class 40 you have there ;)
11:49:07  <FLHerne> Well, yes
11:49:19  <FLHerne> It just looks /slightly/ more different from a 37 :P
11:49:30  <Pikka> I meant D16/2
11:49:37  <FLHerne> Ah
11:49:46  <Pikka> the 10000 was more like a prototype 37 :)
11:49:49  <andythenorth> it doesn't solve my 1975 problem :)
11:50:17  <andythenorth> how many bone-shaped engines can I do :P
11:50:18  <Pikka> why do you need one? what's wrong with the 37? :)
11:50:30  <andythenorth> I want to add an alternative
11:51:07  <andythenorth> I've added a 'joker' roster
11:51:13  <andythenorth> like UKRS 2 extended
11:51:35  <andythenorth> no garratts though :P
11:52:08  <peter1138> hi
11:52:29  <FLHerne> Maybe do the 37 in 1960, then replace it by the 57/0 and add a stretched 67 for express?
11:52:39  <FLHerne> ^in 1990
11:53:33  <andythenorth> could work
11:54:04  <andythenorth> the 57 is really a 50
11:54:06  *** Stimrol has quit IRC
11:54:23  <andythenorth> put diesel engine in this? o_O
11:55:17  *** iSoSyS has joined #openttd
11:55:23  *** bwn has quit IRC
11:57:10  <andythenorth> OR
11:57:32  <andythenorth> I could fix the narrow gauge roster, and give the option of building entirely different railtype
11:57:44  <andythenorth> which might be better than spamming more engines
11:58:18  <peter1138> There was a spate of newgrfs that did that.
11:58:28  <peter1138> Oh yeah, I was scrolled up :p
11:58:44  <andythenorth> 'did that' ? o_O
11:58:46  <andythenorth> did what?
11:59:00  <FLHerne> andythenorth: Isn't that just a class 128?
11:59:12  <peter1138> fake reversing
11:59:41  <peter1138> Hmm, so news and ticker working at normal speed in FFWD...
12:00:34  *** Stimrol has joined #openttd
12:01:02  <peter1138> Heh, viewport scrolling speeds up in FFWD still.
12:06:25  <andythenorth> irish GM bone thing?
12:06:44  <andythenorth> that one's northern irish, to be strictly accurate
12:13:31  *** bwn has joined #openttd
12:20:05  * nielsm thinks about how to modify the music system to support music renderers with pcm output instead of midi
12:20:32  <nielsm> while still keeping midi support in too
12:27:20  <TrueBrain> right .. removed 2 OSes, which were the most annoying :P (PSP and WinCE)
12:27:37  <TrueBrain> others have less of a code impact it seems
12:27:47  <nielsm> wince sounds like it depended on msvc6 support
12:28:08  <TrueBrain> MorphOS and BeOS are next; they also have a few ifdefs around
12:31:31  <frosch123> TrueBrain: what indentation style does source.list use?
12:31:36  <frosch123> you switch from tabs to spaces
12:31:49  <frosch123> in the wince diff
12:32:12  <frosch123> well, i both diffs
12:32:17  <frosch123> +n
12:35:53  <TrueBrain> ugh
12:35:56  <TrueBrain> lemme fix
12:37:14  <TrueBrain> bit annoying how my editor does that
12:38:43  <peter1138> nielsm, music system probably won't care. You'd probably want to modify the mixer to dedicate 2 channels to music.
12:39:00  *** debdog has joined #openttd
12:39:05  <peter1138> Fortunately the mixer is independent (for now) of the sound driver in use.
12:39:11  <TrueBrain> and the PSP diff also had that frosch123 .. should be fixed now, I hope :)
12:39:37  <TrueBrain> nope
12:39:38  <TrueBrain> ffs :P
12:39:49  <nielsm> peter1138 the music playback system would have to know how to handle non-midi files and send them to an appropriate decoder instead
12:39:56  <frosch123> he, i was just questioning my browser cache :p
12:40:03  <nielsm> also, am I reading right that the mixer runs at 11 kHz right now?
12:41:20  <TrueBrain> frosch123: now! :)
12:42:08  <nielsm> oh, I see the sound driver sets mixer rate
12:44:02  <peter1138> non-midi? why?
12:44:30  <nielsm> and for music purposes I think it would make more sense to have a partially separate system, since the mixer currently assumes all sounds are short samples, not continuous streams, so something where the music system gives the mixer a function/object pointer that then supplies samples to get mixed in as well
12:44:36  <nielsm> well, things like mp3 files?
12:44:46  <frosch123> TrueBrain: does jenkins currently build two revisions of the branch in parallel?
12:44:48  <peter1138> Why would you want that?
12:44:49  <nielsm> or an fm synth reading the dos version opl2 music
12:45:20  <peter1138> nielsm, limit your scope.
12:45:24  <peter1138> Stick with MIDI first.
12:45:31  <nielsm> not allowed to dream? :)
12:45:38  <peter1138> Get that working, then consider how to extend to other formats.
12:46:15  <TrueBrain> frosch123: the CI can only do 1 thing at the time, so no?
12:46:20  <TrueBrain> why do you ask?
12:46:49  <frosch123> <- #4 and #5 are active
12:46:53  <TrueBrain> owh, yes
12:46:56  <TrueBrain> the master already picked it up
12:47:03  <TrueBrain> but he cannot assign a node to work on
12:47:26  <TrueBrain> this is because the master has a free slot .. but is not allowed to continue :D
12:47:55  <TrueBrain> and this is because the master spins up a docker, and deligates the full job there .. just I told him he can only have 1 docker slave spinned up :)
12:48:03  <TrueBrain> so master is now waiting for a slave to become available ..
12:48:08  <TrueBrain> guess I could configure it differently, but meh :)
12:58:04  <andythenorth> so if class 15 was (a) not crap (b) not scrapped (c) longer + more powerful?
12:59:09  <andythenorth> looks french :P
12:59:20  *** som89 has joined #openttd
13:11:09  <_dp_> if (t->growth_rate != 0xFFFF) ? t->growth_rate
13:11:21  <_dp_> ha, that's where I fell asleep yesterday xD
13:21:13  <nielsm> listening to ttd dos opl2 music in dosbox while reading the dosbox opl2/3/adlib emulation code
13:21:54  <nielsm> it looks very manageable, and reversing the ttd dos opl2 music driver shouldn't be too bad, a cursory look makes it look very similar to the mpu401 one
13:23:35  <FLHerne> andythenorth: I kinda like it
13:23:36  <frosch123> does it sound any different to the general midi?
13:23:47  <nielsm> well yes
13:23:52  <nielsm> completely different
13:23:55  <FLHerne> Not really a fan of fictional engines, but that's a pretty good one
13:24:40  <nielsm>  <- recordings from a (different) fm synth I made many many years ago
13:25:05  <nielsm> how it sounded when I played it in the mid-90's :)
13:26:25  *** sim-al2 is now known as Guest1248
13:26:27  *** sim-al2 has joined #openttd
13:27:37  *** som89_ has joined #openttd
13:28:03  <peter1138> TTO theme was much better.
13:28:10  <nielsm> it's soothing
13:28:20  <peter1138> Very similar though.
13:28:26  <andythenorth> FLHerne: it's a possibility
13:28:48  <andythenorth> co-co version of this, 1945 or so
13:29:13  <andythenorth> or I make it look more like a 58 and put it in 1975
13:29:22  <andythenorth> the main thing is that engines look different :P
13:30:56  *** Guest1248 has quit IRC
13:31:37  *** som89 has quit IRC
13:35:05  <TrueBrain> wow, Haiku is still an actively maintained OS ..
13:35:39  <TrueBrain> so is MorphOS
13:36:18  <frosch123> i think heise had an headline about a new morphos release last year
13:36:22  <nielsm> well, so are various games dating back 20+ years too :)
13:36:50  <TrueBrain> guess for now I will only be removing BeOS .. as that really is a dead OS :)
13:44:05  <Wolf01> Mmmh, lapdog with XP is just as slow as before... dunno... it was really fast for the first 2 restarts, now it takes again 10-12 minutes to be ready for work
13:46:13  <Alberth> it downloaded old performance characteristics for your type of computer and adaped :p
13:46:20  <Alberth> *adapted
13:46:21  <Wolf01> Probably
13:47:06  <Wolf01> Planned obsolescence
13:47:58  <Alberth> maybe it's trying to download non-existing updates or so
13:48:42  <Wolf01> I think is the antivirus
13:52:57  <TrueBrain> right, enough of removing OSes for now :)
13:56:26  <frosch123> he, there was even a STR_OSNAME_BEOS :p
13:56:47  <TrueBrain> we used to do that for every OS .. but that stopped at some point I guess
13:56:51  <TrueBrain> some even "translated" it :P
13:57:17  <frosch123> sending translator to wiki :)
13:59:15  <peter1138> "All checks have failed"
13:59:30  <frosch123> no idea whether "BEOS_NET_SERVER" would also be used by haiku
13:59:59  <TrueBrain> that is a good question frosch123 .. I assumed it wasn't
14:00:02  <TrueBrain> but that was an assumption
14:00:06  <TrueBrain> cannot really test Haiku ..
14:00:14  <TrueBrain> meh, maybe I should forget about this PR for now
14:00:28  <TrueBrain> first figure out if we want to support Haiku, and if it can be added to the CI someway
14:00:41  <frosch123> meh, more commit hook exceptions
14:02:24  <LordAro> TrueBrain: probably not worth the effort of adding it to CI, although it'd be good to build releases for it
14:02:34  <TrueBrain> LordAro: tomato tomato
14:03:00  <TrueBrain> not saying it should be linked to every PR btw
14:03:06  <TrueBrain> just that once a week orso all targets are tested
14:03:14  <TrueBrain> avoids big surprises whena release approaches :)
14:03:27  <LordAro> yeah, that sort of thing
14:05:55  <frosch123> LordAro: you could also prepare a PR that (1) checks for changed baseset translations, and runs (2)  src/script/api/ and (3) src/script/api/
14:06:24  <frosch123> (1) is something eints currently doesn't, but maybe should?
14:06:36  <frosch123> (2) and in consequence (3) are forgotten all the time
14:08:02  *** snail_UES_ has joined #openttd
14:09:20  <TrueBrain> frosch123: dont forget projects/generate :)
14:09:31  <frosch123> i thought you wanted to remove that :p
14:09:39  <TrueBrain> over time, sure, but that won't be next week :)
14:09:48  <TrueBrain> if cmake works, I expect that we will have both system for a full release
14:12:59  <andythenorth> hmm
14:13:07  * andythenorth trainshedding not bikeshedding
14:13:20  <andythenorth> 100 trains to draw, and I'm obsessing about adding another one, or not :P
14:14:54  <Rubidium> TrueBrain: DOS' latest GCC is 7.3.0, so compiler wise it's definitely new enough... though I think you need to cross-compile from Windows
14:15:11  <TrueBrain> sounds like a challenge
14:15:16  <TrueBrain> how is the endian PR going Rubidium? :)
14:15:53  <Rubidium> LA hasn't made it yet ;(
14:16:04  <TrueBrain> LA?
14:16:41  <Rubidium> LordAro
14:16:46  <TrueBrain> ah
14:16:53  <TrueBrain> you have the diffs ... why not do it yourself?
14:16:57  <TrueBrain> you also have a github account!
14:19:58  <Rubidium> TMWFTLB
14:20:33  <TrueBrain> how rude :(
14:20:40  <TrueBrain> no benefit in making me happy
14:20:42  <TrueBrain> ttssk
14:21:01  <LordAro> :(
14:24:26  *** sim-al2 has quit IRC
14:25:25  *** Gja has joined #openttd
14:28:19  *** Thedarkb1 has joined #openttd
14:36:37  <Rubidium> so I have to recover the account and go through another 20+ steps just for one patch (don't have an intention and/or time to pick further development up in the near future) that make your live easier
14:36:55  <TrueBrain> 20+ steps? Holy crap, what were you planning to do
14:36:57  <TrueBrain> but: yes :D
14:37:02  *** iSoSyS has quit IRC
14:37:26  <TrueBrain> if only you knew your password ... than it would be 2 minutes :P But yes .. now knowing your passwords is annoying :)
14:38:28  *** Wormnest_ has joined #openttd
14:39:23  *** Stimrol has quit IRC
14:39:32  *** Thedarkb1 has quit IRC
14:44:26  *** Thedarkb1 has joined #openttd
14:45:13  <TrueBrain> Rubidium: I guess your patch misses the part for MSVC?
14:45:17  *** Wormnest has quit IRC
14:45:21  <TrueBrain> or do those project files not use it?
14:46:57  *** HerzogDeXtEr1 has quit IRC
14:48:21  *** HerzogDeXtEr has joined #openttd
14:51:41  <TrueBrain> Rubidium: #6762
14:51:55  <TrueBrain>
14:53:14  <Rubidium> TrueBrain: MSVC defined either WIN32 or WIN64, which (before the changes in endian_type) already hard code the endianness
14:53:22  <TrueBrain> k
14:53:23  <TrueBrain> tnx :)
14:53:29  <TrueBrain> removing more fat!
14:54:22  *** iSoSyS has joined #openttd
14:54:46  *** Flygon has quit IRC
14:55:23  *** iSoSyS has quit IRC
15:01:57  *** tokai has joined #openttd
15:01:57  *** ChanServ sets mode: +v tokai
15:08:57  *** tokai|noir has quit IRC
15:22:50  *** tokai|noir has joined #openttd
15:22:50  *** ChanServ sets mode: +v tokai|noir
15:28:06  *** markyisri has joined #openttd
15:28:30  <markyisri> Hello
15:29:06  <TrueBrain> hi
15:29:47  *** tokai has quit IRC
15:30:22  <markyisri> I have a question about the OpenTTD gameplay manual/wiki. Is is possible to download an offline copy?
15:31:01  <LordAro> theres a "Printable version" link on the left side
15:31:20  <LordAro> as for downloading the whole thing though, any sort of web scraper can probably do that for you
15:31:48  <TrueBrain> maybe a better question would be why you want to do that? Wikis tend to be rather large with not-so-useful information?
15:31:56  <_dp_> do people even live offline?
15:32:23  <LordAro> there's an awful lot on the wiki that's out of date
15:32:43  <markyisri> I wanted a comprehensive guide to refer to while learning. I downloaded the game yesterday.
15:32:49  <TrueBrain> any wiki is out-of-date before the save button is pressed the first time LordAro ;)
15:32:56  <LordAro> hehe
15:33:19  <TrueBrain> learn by playing :D But yeah .. I dont really have an answer honestly
15:33:23  <LordAro> markyisri: i'd probably suggest that the only pages you need are Signals and Orders
15:33:33  <LordAro> everything else is fairly intuative for first time
15:34:09  <markyisri> Okay
15:34:11  <markyisri> Thanks for the advice
15:34:48  <LordAro> but good luck! hope you enjoy it :)
15:35:06  <TrueBrain> he will enjoy it! Just the question is .. when will he go to sleep tonight :D
15:35:18  <markyisri> I have enjoyed it already
15:35:36  <markyisri> I was toying around with it while following a 25-minute video tutorial I found on Youtube
15:36:34  *** Gja has joined #openttd
15:37:06  *** markyisri has left #openttd
15:37:38  <TrueBrain> even after 15 years people can surprise you :P
15:37:51  <LordAro> oh?
15:38:01  <TrueBrain> video tutorial on youtube
15:38:08  <TrueBrain> couldn't even predict thatone
15:38:15  <LordAro> oh yes, that's basically how people learn games these days
15:38:19  <LordAro> minecraft started it, i guess
15:44:17  <frosch123> hmm, too late
15:44:36  <frosch123> i wanted to post the only yt signal tutorial where the author actually understood them
15:45:29  <frosch123> <- for reference
15:45:38  *** iSoSyS has joined #openttd
15:45:58  <frosch123> TrueBrain: there are extensive ottd let's plays, even channels dedicated to mostly ottd
15:46:08  <TrueBrain> :o
15:46:42  <frosch123> a few months back some yt douche wanted to impersonate abuse@ and dmca a competitive channel
15:47:01  <TrueBrain> ...... right ...
15:47:06  <TrueBrain> I forgot that people can go really low
15:54:59  *** Cubey has joined #openttd
16:02:53  *** chomwitt has joined #openttd
16:11:51  <peter1138> TrueBrain, yeah, people seem to just watch other people doing stuff these days, instead of just reading.
16:12:07  <TrueBrain> s/reading/player/
16:12:13  <TrueBrain> s/player/playing/
16:12:13  <peter1138> Well that too.
16:12:15  <TrueBrain> :D
16:12:50  <_dp_> You won't get very far in OpenTTD by just playing :p
16:18:24  *** Maraxus has joined #openttd
16:26:52  *** Maraxus has quit IRC
16:33:02  *** glx has joined #openttd
16:33:02  *** ChanServ sets mode: +v glx
16:46:15  <nielsm> frosch123: there you go
16:47:56  <_dp_> tried hard not to change anything with this patch:
16:48:00  <_dp_> kinda failed tho xD
16:48:14  <frosch123> nielsm: yay, now 20 minutes of jenkins :)
16:49:31  <frosch123> well, or 40
16:49:47  <glx> yes the other PR is using the ressources :)
16:49:54  <nielsm> :D
16:50:22  <nielsm> the osx build seems to be significantly slower than the rest, or is it just me?
16:50:50  <__ln__> *resources
17:00:22  <andythenorth> it's slow for me locally
17:00:27  <andythenorth> takes a couple of minutes
17:01:33  <nielsm> I suppose I should do some testing of the non-windows dos music loading code too, from ther other PR
17:01:36  <TrueBrain> Rubidium: merged your endian-removal patches :) Tnx again
17:02:09  <nielsm> since I can just call it from the win32 midi player regardless
17:04:22  <nielsm> first bug found!
17:06:11  <TrueBrain> may it be your last :)
17:06:59  <nielsm> forgot to skip the path separator when generating the temp filename for the .mid file to be written
17:07:21  <nielsm> so it'd have made filenames like 0/GMCAT.mid instead of 0GMCAT.mid
17:08:38  <nielsm> is it reasonable to write those to the content_download/baseset/ directory?
17:11:39  <nielsm> (the unreasonable part might be that the files get re-written each time the song starts, I haven't made a mechanism to check for file already existing, and am not sure how much verification that it's really the right thing would be appropriate)
17:14:21  <frosch123> when ottd downloads a musicset from bananas, it uncompresses the content into a separate folder
17:14:34  <frosch123> i would treat the CAT files like archives
17:15:06  <nielsm> so instead make it content_download/baseset/GM.CAT/0.mid
17:15:08  <frosch123> so uncompresing GM.CAT results in GMCAT/0.mid or somthing
17:15:25  <frosch123> first i wanted to suggest "GM/0.mid", but that would conflict with the win music
17:15:26  <nielsm> and just assume the file is correct if it exists
17:16:10  <frosch123> likely you are also not allowed to name the directory identical to the file, so "GM.CAT/" is also invalid :p
17:16:42  <nielsm> usually GM.CAT would be in baseset/ in the install dir, not in content_download/baseset/
17:16:50  <nielsm> since you don't download the cat files from bananas
17:17:30  <frosch123> true, but people drop their stuff wherever :p
17:18:48  <frosch123> how do you plan to make the .obm?
17:19:11  <frosch123> should the .obm refer to GM.CAT or to GMCAT/ ?
17:19:42  <nielsm> GM.CAT
17:20:02  <nielsm>
17:20:54  <nielsm> I'm not sure if it should be named lowercase or uppercase in the .obm file actually
17:21:38  <frosch123> ""? i doubt it was named like that :p
17:22:26  <nielsm> yes you have to rename it
17:22:37  <nielsm> it's named in both original and deluxe
17:23:11  <frosch123> we also support both dos graphics without needing to rename them
17:23:19  <frosch123> but i see, that would conflict with the unpacking
17:23:26  <nielsm> I also made myself a "original_remix.obm" that uses the theme from TTO but the rest from TTD, and just adds the TTD theme as another song on Ezy Street
17:24:50  <nielsm> but part of what happens is also that the music driver doesn't *have* to write a .mid file to disk to play it, the new win32 driver plays the decoded "mps midi" (as I call it internally) from memory
17:25:38  <TrueBrain> glx: is Cygwin still a thing? Or can we just kill it?
17:26:31  <TrueBrain> nielsm: do you think there is a solution for #6750 ? Or shall we promote it to an issue and leave it for another day to be solved? :)
17:27:18  <TrueBrain> frosch123: nice going with the way you give feedback in PRs. I really like the tone and attention you give to it :) Kudos to you :)
17:27:27  <nielsm> uh the VBS script could be fixed but it'll be bothersome
17:27:42  <glx> cygwin still exist, latest version is from february
17:27:58  <frosch123> TrueBrain: trying my best :p
17:28:11  <nielsm> I entered sysadmin world after powershell became a thing so I'm not at all fluent in vbscript ;)
17:28:16  <frosch123> but actually i just leave the closing to others :p
17:28:39  <TrueBrain> haha :D
17:28:52  <TrueBrain> nielsm: if only spanish is the issue, I am somewhat tempted to just hard-code it :P
17:29:30  <nielsm> oh like, "if filename is spanish.txt write spanish_MX.txt, and if filename is spanish_MX.txt then write spanish.txt" or such?
17:29:33  <nielsm> that's evil
17:30:29  <frosch123> TrueBrain: wasn't the concusion that linux should sort based on basename, without extension?
17:30:42  <TrueBrain> frosch123: I read a lot of weird suggestions :D
17:30:50  <TrueBrain> holy crap, generate.vbs is annoying with all the OK dialogs
17:30:54  <nielsm> I'm actually surprised you have mexican spanish separate, but not brazilian vs european portuguese
17:31:04  <TrueBrain> tnx btw, glx, guess we will leave CygWin alone for now :P
17:31:10  <glx> nielsm: check better
17:31:18  <frosch123> nielsm: they both exist, but use different filenames
17:31:26  <nielsm> TrueBrain, run it from commandline, "cscript generate.vbs"
17:31:34  <frosch123> brazilian_portuguese.txt
17:31:40  <nielsm> WHY
17:31:47  <nielsm> that's just not ISO-like at all!
17:32:06  <glx> we could just rename spanish_MX to mexican_spanish
17:32:32  <frosch123> please don't :p that breaks too much
17:32:42  <TrueBrain> pfff
17:32:45  <TrueBrain> we will overcome :P
17:32:48  <nielsm> or rename spanish.txt to spanish_EU.txt
17:32:53  <TrueBrain> eints just needs a pat on the back :P
17:33:05  <frosch123> i think game scripts are the issue
17:33:21  <TrueBrain> hmm
17:33:22  <glx> scripts use filenames ?
17:33:24  <TrueBrain> alias?
17:33:49  <frosch123> tmwftlb? :p
17:33:54  <TrueBrain> ack
17:34:09  <frosch123> glx: yes
17:35:36  <glx> but openttd itself doesn't care about the filename, that's silly scripts are using it
17:39:12  <frosch123> no ottd does
17:39:39  <frosch123> ottd expects script translations to use the same filename as the ottd translation
17:43:29  <Alberth> no ##grflang  identification in AI/Game scripts
17:48:36  <TrueBrain> nielsm: okay, I think I found a clean way without going to hacking :)
17:48:42  *** som89_ has quit IRC
17:48:49  <TrueBrain> first I tried to make Vbs do the right thing .. but even a manual Sort() gave a different result .. silly Vbs :P
17:48:53  <TrueBrain> and what a horrible scripting language
17:50:33  <frosch123> i would have put the sed script into single quotes
17:50:42  <frosch123> too many $ and ~
17:50:47  <TrueBrain> I just moved code
17:51:02  <TrueBrain> did not want to reinvent too much :)
17:51:17  <TrueBrain> the sed clearly has been working for years, so .. yeah :P
17:51:24  <TrueBrain> (I do agree with you btw :) )
17:51:45  <frosch123> oh, did not notice that it was moved
17:51:54  <TrueBrain> its 2 lines lower! :P
17:51:58  <frosch123> i just assumed that the first sed removed the extension and the second readded it
17:52:25  <TrueBrain> with these things I always try to find the bare minimum to fix it with :)
17:52:27  <TrueBrain> too scred :D
17:52:58  <TrueBrain> funny, 3 languages already updated their strings :P
17:52:59  <nielsm> argh why is FioCreateDirectory marked 'static' in fileio.cpp
17:53:17  <TrueBrain> still surprised by the dedication of our translators
17:53:40  <TrueBrain> nielsm: because nobody ever needed it outside of flileio.cpp? Something easily fixed? :)
17:55:35  <frosch123> one PR referencing another is weird
17:55:40  <TrueBrain> :D:D
17:55:45  <TrueBrain> it made me giggle
17:55:50  <TrueBrain> I could have made a PR to his fork
17:56:04  <TrueBrain> but we should have made an issue out of it first
17:56:09  <TrueBrain> than reference them from PRs
17:56:13  <TrueBrain> GitHub used to do that
17:56:19  <TrueBrain> guess something with overhead
17:56:36  <TrueBrain> I will remove it from my commit message on squashing frosch123 :)
17:58:05  <glx> quickly testing on mingw but it should be ok
17:58:32  <glx> well bash and mingw are never quick ;)
17:58:47  <TrueBrain> yeah .. I had the same issue while testing
17:59:02  <TrueBrain> worse for me was, to test it on linux I first had to drag it in a git on a linux mount (\r\n issues)
17:59:26  <TrueBrain> which was WEIRD
17:59:30  <TrueBrain> but I am happy with WSL
17:59:37  <TrueBrain> lot easier to develop for both OSes
17:59:58  <nielsm> yay now it created content_download/baseset/GM-TTOCAT/0.mid
18:00:01  <nielsm> as intended
18:00:33  <glx> mingw still works as before :)
18:01:23  <TrueBrain> glx: is that a good or bad thing? :)
18:03:41  <frosch123> haha, we have a jenkins queue of over an hour
18:04:22  <TrueBrain> yeah .....
18:04:31  <frosch123> TrueBrain: should we buy another vps with a few more cores?
18:04:36  <TrueBrain> I really either need more CPU power, or make the CI incremental, or install ccache
18:04:54  <TrueBrain> frosch123: if we can afford it :P
18:04:54  <frosch123> i doubt aws would be free, and aws is 3 times more expensive than some vps
18:05:05  <frosch123> (according to my computation last night)
18:05:10  <TrueBrain> we can look at the AWS instances that .. what was it called ...
18:05:22  <nielsm> hm should I put the minor change to fileio.cpp & fileio_func.h in a separate commit or just "hide" it along with the feature code that depends on it?
18:05:30  <TrueBrain> yeah, normal AWS instances are very expensive if you compare them to dedicated servers
18:05:40  <nielsm> AWS you can pay per minute basically
18:05:50  <TrueBrain> and you have this instance type
18:05:51  <nielsm> if you just shut down the instance when you don't need it
18:05:54  <TrueBrain> which is cheap, but not always available
18:06:46  <TrueBrain> Spot
18:06:50  <TrueBrain> AWS Spot instances
18:06:53  <TrueBrain> those are a lot cheaper
18:06:55  <TrueBrain> but might not be available etc
18:07:11  <andythenorth> jenkins has support for AWS spot
18:07:43  <andythenorth> it can stop and start nodes, but you tend to run with a pool of them available AIUI
18:07:50  <andythenorth> it's very cheap, until it's not
18:09:19  <frosch123> TrueBrain: <- something from the C2 series, like C2-60
18:10:06  <frosch123> unless the billing rounds to whole hours :p
18:12:41  <nielsm> here have some more for the jenkins queue
18:23:24  <TrueBrain> okay, math time
18:23:32  <TrueBrain> we had 50 PRs in, what, a month
18:23:39  <TrueBrain> which resulted in, what, 150 CI hits?
18:23:49  <TrueBrain> lets say it doubles
18:24:01  <TrueBrain> means 300 CI hits per month, 10 a day, 1 every 2 hours
18:24:40  <TrueBrain> we currently use 2 cores, build takes, what, 20 minutes .. make that 30
18:24:58  <TrueBrain> so 50% load on a single core
18:25:12  <TrueBrain> lets keep things simple, say we double again, 600 CI hits a month
18:25:17  <TrueBrain> means we need 1 core
18:26:25  <TrueBrain> @calc 0.062 * 24 * 30
18:26:25  <DorpsGek> TrueBrain: 44.64
18:26:38  <TrueBrain> so 45 euro on your URL, for 2 cores
18:26:48  <TrueBrain> @calc 0.062 * 24 * 30 * .121
18:26:48  <DorpsGek> TrueBrain: 5.40144
18:26:51  <TrueBrain> @calc 0.062 * 24 * 30 * 1.21
18:26:51  <DorpsGek> TrueBrain: 54.0144
18:26:53  <TrueBrain> sorry, 55 euro
18:27:00  <frosch123> prices are linear for more cores
18:27:09  <TrueBrain> so, 27 euros
18:27:41  <TrueBrain> for 42 euro a month we can buy a dedicated server similar to what we have now .. with 8 cores
18:27:48  <TrueBrain> just doesnt scale as well (no burst)
18:27:58  <TrueBrain> well, 4 cores, 8 HT
18:28:13  <TrueBrain> so ... 10 euro with no burst vs 27 euros with burst
18:29:46  <TrueBrain> lets see .. AWS spot .. 2 vCPUs, 0.0317 per hour
18:29:57  <TrueBrain> @calc 0.0317 * 24 * 30
18:29:57  <DorpsGek> TrueBrain: 22.824
18:30:06  <TrueBrain> so 22 euro per month, but for 2 cores, so 11 euro
18:30:07  <TrueBrain> with burst
18:30:10  <TrueBrain> but no guarantees
18:30:30  <TrueBrain> linear with CPU too
18:30:33  <frosch123> there are also "no guarantees" on the ovh site
18:30:45  <frosch123> at the bottom
18:30:46  <andythenorth> AWS limit how many spot instances you can buy also
18:30:53  <andythenorth> and then it gets expensive to add more to your account
18:31:12  <TrueBrain> but we only need 4 or so
18:31:15  <andythenorth> not sure how that works, but doubling our Jenkins would cost something like 3x or 4x as much
18:31:17  <andythenorth> at work
18:31:26  <andythenorth> I am hazy on the details :P
18:31:29  <TrueBrain> :D
18:31:41  <TrueBrain> but EC2 Spot is the cheapest solution for us I guess
18:31:43  <andythenorth> anyway, I can sponsor x amount per month
18:31:52  <andythenorth> I used to sponsor coop, but that seems to not be a thing
18:32:04  <frosch123> @calc 300*40/60*0.01
18:32:04  <DorpsGek> frosch123: 2
18:32:05  <TrueBrain> would mean we have to pay AWS around 20 euro per month for it, give or take a few
18:32:21  <michi_cc> For the OVH cloud I read lots of 'Billed to the nearest hour'.
18:32:32  <andythenorth> I can do like £15 / month
18:32:35  <TrueBrain> they almost always are billed to the next hour, honestly :)
18:32:40  <andythenorth> how much do we have orudge? o_O
18:32:41  <frosch123> 2 euro per month :p
18:32:52  <frosch123> i have no idea what this "no guarantee" means :p
18:32:57  <TrueBrain> andythenorth: we havent had a fundraiser in years; nuff said
18:33:10  <TrueBrain> frosch123: if you dont say what you calculate, it means very little to me :)
18:33:18  <andythenorth> frosch123: 'no guarantee' _usually_ means contended instances
18:33:24  <andythenorth> and they can go away at no notice
18:33:40  <frosch123> TrueBrain: you said 300 hits per month, 40 cpu minutes per hit
18:33:58  <frosch123> @calc 300*40/60*0.085/2
18:33:58  <DorpsGek> frosch123: 8.5
18:34:20  <frosch123> i don't get your 27 euros
18:34:36  <frosch123> though i skipped VAT
18:34:41  <TrueBrain> I calculated 50% load on a single core, with 300 hits per month, and say 30 minutes per hit
18:35:03  <TrueBrain> so to make calculations easier, as round to an hour will fuck us over
18:35:09  <TrueBrain> say 100% load on a single core
18:35:19  <TrueBrain> means if you look at the price for a single core, it will always be under, never over
18:35:33  <TrueBrain> they ask 0.062 euro per hour for 2 cores
18:35:49  <TrueBrain> @calc 0.062 * 24 * 30 * 1.21
18:35:49  <DorpsGek> TrueBrain: 54.0144
18:35:56  <TrueBrain> 24 hours in a day, 30 days in a month, + VAT
18:36:34  <frosch123> but that is for constant usage
18:36:42  <TrueBrain> 1 CPU core at 100% load
18:36:44  <TrueBrain> is pretty constant yes
18:36:53  <TrueBrain> that was the whole point of picking that :)
18:36:56  <frosch123> @calc 0.062 * 20/60 * 300 * 1.21
18:36:57  <DorpsGek> frosch123: 7.502
18:37:12  <frosch123> your computations are far away from the 300 hits
18:37:20  <TrueBrain> yes, it assumes we will have 4 times more PR than we have now, but .. with the rounding of the hours, it most likely doesnt matter if we have 300 CI hits or 1200
18:37:35  <TrueBrain> yes ... I said that
18:37:44  <TrueBrain> 300 hits with 30 minutes is ~50% .. I doubled that
18:38:07  <andythenorth> if it runs out of nodes stuff will just queue up no?
18:38:08  <TrueBrain> these "per hour" prices basically means: every CI hit will cost an hour
18:38:38  <TrueBrain> so if we keep the 300, it would mean
18:38:43  <TrueBrain> @calc 0.062 * 300 * 1.21
18:38:43  <DorpsGek> TrueBrain: 22.506
18:38:48  <TrueBrain> 22 euros
18:39:05  <TrueBrain> either way, the numbers were not that relevant to me, more how they compare to each other
18:39:13  <TrueBrain> basically, OVH cloud is expensive, as is EC2
18:39:21  <TrueBrain> EC2 spot is cheap, about as expensive as our own dedicated server
18:39:25  <TrueBrain> EC2 spot can only scale
18:39:34  <TrueBrain> so it is a balance game; what do we want
18:39:40  <TrueBrain> consistant throughput, or burst
18:39:53  <frosch123> to me it looks we need burst on weekend
18:40:04  <TrueBrain> so EC2 Spot is a good candidate in that case
18:40:08  <frosch123> which may mean that we can go for the SANDBOX thing, when no business uses them
18:40:27  <TrueBrain> sandbox?
18:40:40  <frosch123> <- at the bottom
18:40:43  <TrueBrain> ah
18:40:51  <frosch123> 1/8 the price, but no guarantees
18:40:55  <frosch123> whatever that means :p
18:41:05  <TrueBrain> AWS does the same .. but I guess AWS is a bit bigger :P
18:41:11  <TrueBrain> higher chance they are available :D
18:41:30  <TrueBrain> I think 2 GB is not enough btw .. had issues when the VM was running on 2GB
18:42:02  <TrueBrain> wow, Windows is expensie :o
18:42:39  <TrueBrain> "from the three available" continues to list 4
18:42:40  <TrueBrain> lol
18:42:50  <frosch123> 60% more or so
18:43:24  <TrueBrain> no SANDBOX for Windows :P
18:43:41  <TrueBrain> and almost double for the C2-7 for example
18:44:15  <TrueBrain> either way, these things, are really a matter of trying
18:44:20  <TrueBrain> no clue how they work and how well they work
18:44:35  <TrueBrain> and the other approach is simply adding ccache for the linux CIs :P
18:44:43  <TrueBrain> would drop their times drasticly
18:44:56  <TrueBrain> just shit-annoyed AWS never replied :(
18:46:09  <TrueBrain> glx: did mingw work with my generate fix? (did it produce no diff on my branch)?
18:48:43  *** Pikka has quit IRC
18:50:22  *** gelignite has quit IRC
18:52:12  *** ToBeFree has joined #openttd
18:58:19  <glx> TrueBrain: yes it works, still produces the same diff as generate.vbs
18:58:46  <glx> hence the work as before ;)
18:58:58  *** ToBeFree has quit IRC
18:59:17  <TrueBrain> cool!
19:13:42  *** Supercheese has joined #openttd
19:18:35  *** sla_ro|master has quit IRC
19:31:36  *** cbrowne has joined #openttd
19:32:52  <TrueBrain> @calc 0.062 / 2 * 1.21
19:32:53  <DorpsGek> TrueBrain: 0.03751
19:33:13  <TrueBrain> @calc 0.0317 / 2
19:33:13  <DorpsGek> TrueBrain: 0.01585
19:33:48  <TrueBrain> @calc 0.014 / 2
19:33:48  <DorpsGek> TrueBrain: 0.007
19:34:12  <TrueBrain> frosch123: another way of saying this, the above are the prices per CI run :D (OVH Cloud, AWS Spot, OVH SANDBOX)
19:34:25  <TrueBrain> so we can charge people to create their PR to cover it :D
19:34:41  <LordAro> in USD?
19:35:00  <frosch123> ok, then i only found the wrong aws stuff
19:35:20  <LordAro> OVH has sponsored OTTD in the past...
19:35:51  <frosch123> sponsoring does not work if you have noone to contantly run after them
19:35:55  <TrueBrain> EC2 spot
19:36:04  <frosch123> at which point you have to compare it to your salary :p
19:36:47  <TrueBrain> only works for jenkins stuff .. not for hosting etc
19:40:25  <frosch123> fpga instances? who uses that?
19:40:33  <frosch123> people who want to decrypt stuff?
19:45:40  <TrueBrain> any specific job I guess ..
19:47:14  <andythenorth> sponsoring bla
19:47:15  <andythenorth> h blah
19:47:26  <andythenorth> I was paying £15 / month for world of tanks
19:47:33  <andythenorth> and it's not even pay-to-win :P
19:47:39  <LordAro> andythenorth: oh dear
19:47:41  <andythenorth> I can cover some cloud hosting
19:47:46  <LordAro> you're one of *those* people
19:48:43  <andythenorth> I quit WoT
19:51:44  <TrueBrain> I thought you paid us with your grumpyness every day andy?
19:52:30  <andythenorth> WoT made me grumpy :P
19:54:59  <FLHerne> andythenorth: Wow, is production in FIRS 3 supposed to be this high?!
19:55:22  <FLHerne> My brand-new, unfed Orchard & Piggery produces 185t of fruit per month...
19:55:52  <andythenorth> yea
19:56:00  <nielsm> FLHerne on the other hand, it can't ever increase to 2000 t/month
19:56:06  <andythenorth> although you got lucky with randomisation there
19:56:22  <FLHerne> The other ones are all >100
19:56:36  <FLHerne> This is going to be a pain with my little 15t steam lorries :P
19:57:17  <nielsm> part of the challenge with FIRS is that you basically have to server lots and lots of industries to get large amounts of raw materials, and also send supplies back
19:57:56  <andythenorth> depends on the economy
19:57:57  <FLHerne> nielsm: Yes, I've played with FIRS before, just not this version
19:57:57  <andythenorth> but yeah
19:58:09  <andythenorth> previously farms were smaller and clustered densely
19:58:26  <andythenorth> in 3 they are larger (higher production) and cluster less densely
19:58:39  <FLHerne> In the older one, initial production is 3-4 dozen, so you really have to use supplies to get a decent output
19:59:32  <FLHerne> I can't imagine ever /wanting/ to feed supplies to these, they'll be enough of a pain to serve as it is
20:00:42  <FLHerne> Hm, wasn't there a parameter for production level once? 'Generous' etc. ?
20:00:50  <FLHerne> Or is my memory playing tricks?
20:01:54  <andythenorth> there's a parameter for effect of supplies
20:02:26  <andythenorth> most FIRS primaries can be served by two trains around TL5
20:02:40  <andythenorth> some get randomised up to 400 units or so / month
20:02:47  <andythenorth> which needs TL8 or 3 trains
20:02:59  <FLHerne> Aargh, even the original industry has silly production levels
20:03:03  <FLHerne> *economy
20:03:14  <FLHerne> Well, at least it's open-source
20:03:20  <andythenorth> it's easy to adjust
20:03:57  <andythenorth> the net production is about the same as you'd get previously from 3 clustered farms
20:04:11  <andythenorth> but now there is less yak-shaving to collect it
20:04:22  <FLHerne> Yeah, where's the fun in that? :P
20:05:28  <andythenorth> there's a reason FIRS 2 has a different grfid :P
20:05:33  <andythenorth> still on bananas eh
20:06:00  <FLHerne> That's a point, might be easier
20:06:10  <FLHerne> OTOH, I do already have a FIRS checkout
20:08:28  <andythenorth> ha ha
20:08:29  <andythenorth>
20:08:36  <andythenorth> that's the Little Bear engine in Horse 1
20:08:43  <andythenorth> except we completely invented the Little Beaer
20:09:49  <FLHerne> Blegh
20:10:10  <FLHerne> New firs doesn't like being built in a git-hg checkout
20:10:23  *** synchris has quit IRC
20:10:26  <andythenorth> unfortunate
20:10:34  <andythenorth> I'd switch to git...but eh :(
20:13:37  <FLHerne> Meh, I can hack bin/hg-info to return nonsense :P
20:14:56  <FLHerne> Oh, I don't even need to do that
20:19:21  * FLHerne sets REPO_VERSION to 9999, wcgw
20:23:48  *** Alberth has left #openttd
20:29:12  *** cbrowne has quit IRC
20:33:14  <peter1138> andythenorth, DO IT
20:37:54  *** frosch123 has quit IRC
20:38:46  *** iSoSyS has quit IRC
20:38:54  * andythenorth also bye
20:38:54  *** andythenorth has left #openttd
21:20:57  *** nielsm has quit IRC
21:23:14  *** sim-al2 has joined #openttd
21:31:33  *** Stimrol has joined #openttd
21:31:47  *** som89_ has joined #openttd
21:39:42  *** Stimrol has quit IRC
21:40:33  *** Stimrol has joined #openttd
21:45:05  *** som89 has joined #openttd
21:47:30  <Eddi|zuHause> anyone here have any clue how to diagnose windows performance problems? this system is really really slow, says disk I/O is at 100%, but only like 1MB/s
21:48:52  *** som89_ has quit IRC
21:49:32  *** Wormnest_ has quit IRC
22:00:16  *** som89 has quit IRC
22:03:25  *** beno has joined #openttd
22:03:33  *** som89 has joined #openttd
22:04:22  *** Supercheese has quit IRC
22:04:39  *** Supercheese has joined #openttd
22:07:50  *** som89_ has joined #openttd
22:10:07  *** Thedarkb1 has quit IRC
22:12:17  *** som89 has quit IRC
22:22:03  *** Stimrol has quit IRC
22:24:43  *** Progman has quit IRC
22:26:43  *** Gja has quit IRC
22:30:33  *** Stimrol has joined #openttd
22:32:35  *** Stimrol has quit IRC
22:33:59  *** Thedarkb-X40 has joined #openttd
22:40:10  *** Stimrol has joined #openttd
22:42:07  *** beno has quit IRC
23:37:22  *** FLHerne has quit IRC
23:37:37  *** Eddi|zuHause has quit IRC
23:39:34  *** Eddi|zuHause has joined #openttd
23:40:31  <Eddi|zuHause> hm, i sorta isolated it to "tile data model server" service... but 1) no tiles should be active, and 2) it restarts immediately when i stop it
23:49:20  <Wolf01> uhm
23:49:28  <Wolf01> 'night
23:49:31  *** Wolf01 has quit IRC

Powered by YARRSTE version: svn-trunk