IRC logs for #openttd on OFTC at 2021-11-02
⏴ go to previous day
00:03:19 *** Kitrana has quit IRC (Read error: Connection reset by peer)
00:10:51 *** nielsm has quit IRC (Ping timeout: 480 seconds)
00:12:09 *** WormnestAndroid has joined #openttd
00:19:34 *** Kitrana has joined #openttd
00:22:03 *** Kitrana1 has joined #openttd
00:27:35 *** Kitrana has quit IRC (Ping timeout: 480 seconds)
00:36:40 *** Kitrana1 has quit IRC (Read error: Connection reset by peer)
00:39:43 *** Kitrana has joined #openttd
00:42:08 *** Kitrana1 has joined #openttd
00:47:45 *** Kitrana has quit IRC (Ping timeout: 480 seconds)
01:27:45 *** tokai|noir has joined #openttd
01:27:45 *** ChanServ sets mode: +v tokai|noir
01:34:35 *** tokai has quit IRC (Ping timeout: 480 seconds)
01:53:07 *** snail_UES_ has joined #openttd
02:38:32 *** Wormnest has quit IRC (Quit: Leaving)
03:34:30 *** felix has quit IRC (Ping timeout: 480 seconds)
04:06:17 *** snail_UES_ has quit IRC (Quit: snail_UES_)
04:31:49 *** ChanServ sets mode: +v tokai
04:38:46 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
06:37:19 *** sla_ro|master has joined #openttd
07:36:35 *** tokai|noir has joined #openttd
07:36:35 *** ChanServ sets mode: +v tokai|noir
07:43:35 *** tokai has quit IRC (Ping timeout: 480 seconds)
07:44:05 *** andythenorth has joined #openttd
07:46:58 *** juanma has quit IRC (Quit: Leaving)
07:55:16 *** WormnestAndroid has quit IRC (Remote host closed the connection)
09:39:51 *** andythenorth has quit IRC (Quit: andythenorth)
10:14:55 *** Etua has quit IRC (Quit: Etua)
11:56:15 *** Kitrana1 has quit IRC (Quit: Leaving.)
11:56:37 *** WormnestAndroid has joined #openttd
12:21:10 <DorpsGek> [OpenTTD/OpenTTD] Siob21 commented on issue #9664: [Crash]: Whenever I idle on the main menu for too long the game gives a fatal error https://git.io/JPRBE
12:53:35 *** Kitrana has joined #openttd
13:53:05 *** ChanServ sets mode: +v tokai
14:00:00 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
15:08:49 *** Wormnest has joined #openttd
16:34:14 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:55:17 *** tokai|noir has joined #openttd
16:55:17 *** ChanServ sets mode: +v tokai|noir
17:02:20 *** tokai has quit IRC (Ping timeout: 480 seconds)
17:36:13 *** andythenorth has joined #openttd
17:36:23 <andythenorth> so we have m1 native support yes? :P
17:36:56 <glx> main issue currently is no windows builds at all
17:37:38 <glx> so we can't release 12.1 and stop the flow of duplicate bug reports ;)
17:44:27 *** frosch123 has joined #openttd
17:49:43 <andythenorth> if this migration ever completes, I will see how fast FIRS compiles on m1 :P
17:50:09 <andythenorth> the indicative python benchmarks I've seen aren't any faster on m1 for some cases
17:50:14 <andythenorth> and are insanely faster for some other cases
18:05:32 *** frosch12 has joined #openttd
18:09:30 *** frosch123 has quit IRC (Ping timeout: 480 seconds)
18:23:15 <glx> copy_sprite_layout still not fixed, but it should be easier to do now
18:31:38 *** HerzogDeXtEr has joined #openttd
18:40:36 *** gelignite has joined #openttd
18:42:45 <glx> and of course nothing forces property block to be defined before graphics block
18:49:38 <DorpsGek> - Update: Translations from eints (by translators)
18:57:53 *** andythenorth has quit IRC (Quit: andythenorth)
18:58:23 *** andythenorth has joined #openttd
19:01:43 *** andythenorth has quit IRC ()
19:13:07 *** andythenorth has joined #openttd
19:33:34 *** andythenorth has quit IRC (Quit: andythenorth)
19:34:19 *** andythenorth has joined #openttd
19:42:41 *** andythenorth has quit IRC (Quit: andythenorth)
19:47:35 <frosch12> why is the suggest-code-change thing so hidden in gh...
19:47:43 <frosch12> every time i have to search 5 minutes
19:48:17 *** andythenorth has joined #openttd
19:48:24 *** andythenorth has quit IRC ()
19:57:53 *** andythenorth has joined #openttd
20:00:16 *** andythenorth has quit IRC ()
20:11:04 <frosch12> checked the easy parts :p
20:11:17 <frosch12> spritelayout/foundations are for another day
20:22:46 <glx> I'm still touching spritelayout stuff
21:07:41 *** andythenorth has joined #openttd
21:18:49 *** andythenorth_ has joined #openttd
21:18:49 *** andythenorth has quit IRC (Read error: Connection reset by peer)
21:19:15 *** gelignite has quit IRC (Quit: Stay safe!)
21:22:06 <andythenorth_> yeah python is twice as slow on m1
21:22:18 <andythenorth_> that's an interesting result
21:22:54 <andythenorth_> well more accurately, Iron Horse takes twice as long to compile
21:26:43 <andythenorth_> OpenTTD is not particularly fast either
21:28:02 <frosch12> are you running the x64 binary in the emulator? :p
21:28:38 <andythenorth_> don't we have a universal binary?
21:29:21 <andythenorth_> it's flagged to run for Apple Silicon
21:30:57 <andythenorth_> seems I turned off hardware acceleration as that was slower on intel macs
21:31:33 <andythenorth_> yeah that's pretty slow
21:32:48 <EER> I wonder if the rosetta version could be faster? Depending on how optimized python is for m1/arm ... hmm :)
21:33:33 <andythenorth_> python isn't relevant to openttd speed, but is for nml compiles :)
21:33:54 <andythenorth_> and nml compiles are horribly horribly slower on m1
21:34:26 <andythenorth_> openttd is putting out about 39x on ffwd where my i9 puts out about 13x for the same savegame
21:34:32 <andythenorth_> not particularly impressive
21:34:58 <EER> are those numbers reversed? Else I'd say x3 would be impressive?
21:35:43 <andythenorth_> the £999 macbook air already does x3
21:36:09 <andythenorth_> this is a £2.5k laptop, the numbers are not good for that much spend
21:36:30 <glx> but you have an apple on it ;)
21:36:52 <andythenorth_> I already had one, I only bought a new one to see if it's faster :P
21:37:01 <EER> Is it one of the new ones? max/pro? (not that that will help, I think python is mostly single threaded)
21:37:09 <andythenorth_> OpenTTD mac compile is also still slow
21:37:27 <andythenorth_> it's over 1 minute still
21:37:57 <frosch12> really? are you passing -j ?
21:38:17 <andythenorth_> I gave it -j 19
21:38:37 <andythenorth_> not sure what's ideal, it has 10 cores
21:38:48 <andythenorth_> previous I had 8 cores, and 19 was about right
21:42:43 <andythenorth_> i9 compile was 2m 50s, m1 is 1m 26s
21:44:33 <frosch12> debug build takes 55s on 8 core/16 thread (-j16) ryzen7
21:45:02 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:46:37 <andythenorth_> slow apple is slow
21:48:37 *** sla_ro|master has quit IRC ()
21:53:17 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:56:40 <andythenorth_> hmm python 3.10 is m1 native
21:56:50 <andythenorth_> that makes a difference
21:58:02 *** Samu has quit IRC (Quit: Leaving)
21:59:34 *** andythenorth_ has quit IRC (Read error: Connection reset by peer)
22:00:16 <michi_cc> I don't really like the treatment of CommandCallbacks, but that's the only good idea I had to solve for how AI/GS and network together could be handled.
22:00:18 *** andythenorth has joined #openttd
22:00:48 <glx> funny how only the second + is transformed to %2B
22:03:50 * andythenorth rebuilt nml with python 3.10
22:04:03 <andythenorth> performance is about equivalent m1 <> i9
22:06:43 *** andythenorth_ has joined #openttd
22:06:43 *** andythenorth has quit IRC (Read error: Connection reset by peer)
22:10:51 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:11:49 *** EER has quit IRC (Remote host closed the connection)
22:27:10 <dP> I still like my command api more :p
22:28:01 <dP> and CommandCallbacks are kinda awkward in general, it may be better to just embed them into the command itself
22:33:44 <michi_cc> AI/GS is the problem there as the scripting system uses a callback for all command calls. Right now this is implicitly resolved by the callback address matching, but you could possibly also hard-code that it is an AI/GS command on the network protocol to resolve the callback.
22:42:47 <dP> another option would be to add some kind of command id and have general client-side callbacks based on that
22:43:08 *** andythenorth_ has quit IRC (Quit: andythenorth_)
22:43:09 <dP> that's kind of what I do there with queue_with_callback in my snippet actually
22:45:05 <dP> since I can't add CommandCallback while keeping network compatibility
22:47:12 <frosch12> hmm, i never knew what the "P" in DoCommandP stands for... but "post" is actually plausible
22:48:11 <michi_cc> I have no idea what the idea behind the P is either, but Post seemed good enough for me :)
22:54:14 <frosch12> "post" matches the naming schema c++2x executors, so it's good
22:56:13 <glx> only issue I can see in my station implementation is if property block is after graphics block, as prop 08 will be after prop 0A or prop 1A in this case
22:57:51 <dP> which openttd should totally do imo, they're awesome
22:58:21 <dP> but here for example rails are built in the callback so doing the preview requires complete code duplication
22:58:45 *** andythenorth has joined #openttd
22:58:56 <dP> when ideally tool should construct some kind of command container, show preview based on test run and then just send the command when user builds it
23:02:59 *** frosch12 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
23:16:01 *** andythenorth_ has joined #openttd
23:21:05 *** andythenorth has quit IRC (Ping timeout: 480 seconds)
23:25:26 <peter1138> Should I Windows 11?
23:26:41 <peter1138> Huh, Android apps... why...
23:26:58 *** andythenorth_ has quit IRC (Quit: andythenorth_)
23:27:19 *** andythenorth has joined #openttd
23:43:12 * andythenorth lost in which python modules are slow :P
23:47:36 *** snail_UES_ has joined #openttd
23:50:58 *** andythenorth has quit IRC (Quit: andythenorth)
continue to next day ⏵
