IRC logs for #openttd on OFTC at 2019-01-13
⏴ go to previous day
00:00:51 <Eddi|zuHause> andythenorth: so that definitely needs more work
00:00:52 <andythenorth> 2019-01-12T22:39:23.2400344Z Running ai/regression/tst_regression... failed!
00:01:15 <Eddi|zuHause> andythenorth: the linux ones contain actual regression output
00:01:39 <frosch123> andythenorth: the window still scrolls when holding the button down
00:02:19 <frosch123> i have no idea whether parts of that diff were committed or whether someone else implemented the same
00:02:56 <frosch123> (svn r27820) -Fix: Allow dropdown to be drawn above origin widget even with scrollbar (Juanjo)
00:03:43 <andythenorth> I can't trigger any scroll behaviour that looks weird
00:03:58 <frosch123> ah, that's the reivision that's mentioned in the issue as well
00:04:36 <frosch123> andythenorth: make the combobox open above the widget with a scrollbar
00:04:56 <frosch123> open the combobox with keeping the button pressed down, not releasing it
00:04:58 <andythenorth> frosch123: I did
00:05:08 <andythenorth> then I can mousewheel scroll as expected
00:05:21 <Samu> assert((dist <= rev_dist) == (dist2 <= rev_dist2));
00:05:26 <frosch123> there are two ways to use a combobox: 1 click open + 1 click select, and press to open + release to select
00:05:32 <glx> <Eddi|zuHause> andythenorth: the linux ones contain actual regression output <-- indeed in the windows vbs I just do a <content of file A> = <content of file B> with no diff
00:05:53 <Samu> yeah, this reversal is bad behaviour, well uhm, how am i gonna do this
00:05:55 <frosch123> in the second method the window scrolls with the mouse is above/below the combowindow
00:06:07 <glx> that's assuming the sh/awk fails too on other platforms
00:06:24 <andythenorth> but I have the ottd mac-specific scroll stuff disabled, because it's bollocks
00:06:27 <andythenorth> let me turn on it
00:06:43 <frosch123> well, my repository is in november, no idea whether something changed in past 2 months
00:07:08 <Eddi|zuHause> glx: i can't tell from the regression output what it was trying to do and whether that is a "good" or "bad" change
00:09:08 <glx> it's related to depots ;)
00:09:23 <Eddi|zuHause> glx: yeah, i see that :p
00:09:28 <LordAro> i don't actually know how to get larger interface on the menu screen
00:09:39 <Eddi|zuHause> glx: whic might be somewhat expected since we changed depot code :p
00:09:45 <LordAro> which makes it a bit difficult to test #6212
00:09:52 <glx> and line numbers should match the .nut
00:11:22 <glx> ah no it doesn't because iterators
00:11:26 <andythenorth> LordAro: game options?
00:11:40 <LordAro> how long has that been there>
00:12:14 <andythenorth> since peter1138 did it :)
00:12:21 <andythenorth> best feature ever, at the time
00:12:43 <Eddi|zuHause> we need 1.5 zoom. or sqrt(2)
00:13:31 <glx> Eddi|zuHause: running regression with -k flag used to keep result.txt, but I think it's overwritten by the second test
00:13:31 <nnyby> C++ is so much harder to use than javascript/python lol
00:14:02 <Eddi|zuHause> glx: not going to look at this today...
00:14:21 <andythenorth> because votes = patches, right? :P
00:14:37 <frosch123> andythenorth: so, something changed from november to now
00:14:52 <frosch123> someone broke or removed the scrolling
00:15:11 <glx> Eddi|zuHause: I'll look at it and put it in comments
00:17:26 <andythenorth> I'm assuming it's somewhere else :P
00:17:41 <frosch123> i am compiling the revision before the ontick stuff
00:19:24 <LordAro> 6266 looks fine though
00:19:31 <LordAro> i'll make that into a PR at some point, probably
00:19:57 <andythenorth> cleared out a bunch, saved a few
00:20:12 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7044: Codechange: Define INVALID_TOWN as a TownID. https://git.io/fhnW3
00:20:14 <frosch123> so, the scrolling still happens before the ontick stuff, and probably again with 7048
00:20:38 <andythenorth> frosch123: shall I update the issue, or you will?
00:22:58 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #7048: Fix #6780: Some windows didn't get updated from OnTick() https://git.io/fhnB7
00:23:15 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7048: Fix #6780: Some windows didn't get updated from OnTick() https://git.io/fhnzB
00:23:25 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7048: Fix #6780: Some windows didn't get updated from OnTick() https://git.io/fhnBV
00:23:35 <LordAro> andythenorth: there you go :p
00:26:06 <andythenorth> frosch123: repro-ed
00:26:16 <glx> oh macOS task is the best way to look at regression diff ;)
00:26:33 <glx> in linux tasks it's hidden in the logs
00:27:31 <glx> but still doesn't tell what the regression was testing exactly
00:28:49 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6060: Allow drawing dropdown lists with scrollbars above the widgets https://git.io/fhnzu
00:29:58 <andythenorth> but very bedtime
00:33:17 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6078: Orders: it does not jump order by remaining lifetime https://git.io/fhnzr
00:34:42 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #5321: AI: allow placing waypoints like human player (next to each, other, covering more than 1 tile etc) https://git.io/fhnz6
00:46:55 *** rocky1138 has joined #openttd
00:58:59 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhnzj
00:59:02 <glx> ok I have a PR to write to improve regression
01:17:58 <Samu> *fstd.best_track = INVALID_TRACK; expression must have class type
01:19:53 <Samu> fstd->best_bird_dist = INVALID_TRACK;
01:23:59 <glx> it's a fun regression error
01:40:14 <glx> ok found the cause, thanks to regression :)
01:54:33 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z opened pull request #7052: Change: Allow drawing dropdown lists with scrollbars above the widget… https://git.io/fhngQ
02:03:06 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 opened pull request #7053: Fix: make -k switch useful when testing regression https://git.io/fhngj
02:07:36 *** Thedarkb1-T60 has joined #openttd
02:09:50 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7052: Change: Allow drawing dropdown lists with scrollbars above the widget… https://git.io/fhn2f
02:36:01 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z closed pull request #7052: Change: Allow drawing dropdown lists with scrollbars above the widget… https://git.io/fhngQ
02:36:02 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7052: Change: Allow drawing dropdown lists with scrollbars above the widget… https://git.io/fhn2s
03:04:57 <Eddi|zuHause> " assert_compile(VEH_TRAIN == (int)ScriptTile::TRANSPORT_RAIL);" is that really necessary?
03:05:56 <Eddi|zuHause> well, i can leave it in
03:08:44 <glx> unless enum TransportType and enum VehicleType equivalence is checked somewhere else it's better to keep it
03:11:04 <Eddi|zuHause> still leaves the "unclean" part of the queue to be resolved
03:11:04 <glx> I'm doing a quick fix for 7050
03:14:42 <glx> hmm mingw decided to fail to build
03:15:12 <glx> oh I know why, the c++11 PR has been merged
03:15:57 <Eddi|zuHause> can't help you with that :p
03:17:07 <Samu> r can generate a number from 0 to 255, isn't it?
03:17:37 <Samu> what if it is 30 and then r-30 happens?
03:20:45 <Samu> r 4294967284 unsigned int
03:20:54 <Samu> oh, it goes back to a big number?
03:22:44 <DorpsGek_II> [OpenTTD/OpenTTD] btzy commented on pull request #7005: Fix #7004: Redraw linkgraph overlay correctly after zoom https://git.io/fhn27
03:27:59 <Samu> I don't think 4294967284 <= 127
03:28:20 <Samu> is this a bug? intended bug?
03:30:30 <DorpsGek_II> [OpenTTD/OpenTTD] btzy commented on pull request #7040: Fix: Round up deltas for smooth scrolling, so that target will eventually be reached https://git.io/fhn2A
03:31:05 <DorpsGek_II> [OpenTTD/OpenTTD] btzy updated pull request #7040: Fix: Round up deltas for smooth scrolling, so that target will eventually be reached https://git.io/fhnn8
03:33:15 <DorpsGek_II> [OpenTTD/OpenTTD] btzy updated pull request #7040: Fix: Round up deltas for smooth scrolling, so that target will eventually be reached https://git.io/fhnn8
04:07:38 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick updated pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhZzm
04:24:17 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhnaC
05:54:04 *** HerzogDeXtEr1 has joined #openttd
05:55:02 *** Thedarkb1-T60 has joined #openttd
06:57:39 *** Alberth has joined #openttd
06:57:40 *** ChanServ sets mode: +o Alberth
07:18:15 *** triolus has joined #openttd
07:21:48 *** triolus has joined #openttd
08:16:23 *** triolus has joined #openttd
08:18:35 *** triolus has joined #openttd
08:27:57 *** stefino has joined #openttd
08:29:28 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6060: Allow drawing dropdown lists with scrollbars above the widgets https://git.io/fhnwt
08:30:55 *** andythenorth has joined #openttd
08:36:47 <andythenorth> it might be a tile check that prevents land
08:40:07 <stefino> mmm so the code itself like "spec_flags: bitmask(IND_FLAG_BUILT_ON_WATER);" doesnt mean that the industry IS on the water but it CAN be built on the water?
08:40:47 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7053: Fix: make -k switch useful when testing regression https://git.io/fhnwG
08:40:56 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7053: Fix: make -k switch useful when testing regression https://git.io/fhngj
08:40:59 <andythenorth> stefino: yes, that permits water, but doesn't enforce water
08:41:39 <andythenorth> LordAro: early to rise, late to bed :o
08:44:02 <LordAro> andythenorth: i wanted to go cycling this morning, but the weather isn't looking great
08:44:16 <stefino> okay :) second problem is that I code post office ..( it is the same code for all industries cause I started yesterady. So diferences are only in flags or output, input cargo ) https://paste.openttdcoop.org/pthph0txs but this post office accept and produce coal :D instead mail
08:44:30 * andythenorth just built a fishing hole
08:45:16 * andythenorth just built another fishing hole, inland, at the top of a mountain
08:46:12 <andythenorth> stefino: try setting 'override' property too
08:46:18 <andythenorth> this came up for nielsm recently
08:46:30 <andythenorth> there might be undocumented behaviour with 'substitute' and 'override'
08:46:44 <DorpsGek_II> [OpenTTD/OpenTTD] Alberth289346 opened pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnwC
08:47:22 <andythenorth> if I remember correctly, you're getting coal because something from default coal isn't being over-ridden correctly
08:50:02 <stefino> but post office is not in the old game so I cant to use override or substite
08:50:44 *** sla_ro|master has joined #openttd
08:52:50 <Alberth> sounds like not using cargo labels?
08:53:32 <Alberth> nice inland water industries andy :)
08:53:59 <Alberth> sorry for the PS, but it seemed too simple not to take it :)
08:57:15 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnwR
08:57:35 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro closed issue #6408: Signals: confusing concept (wording): "Signal density" https://git.io/fhnw0
08:57:36 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnwC
09:01:36 <stefino> this is all code with post_office (I delete another industries for this sample)
09:02:44 <andythenorth> Alberth: PRs are good :)
09:03:18 <stefino> I did it with YETI instructions in devzone. I can't say that I am progammer :D
09:03:25 <Alberth> also a way to close issues :)
09:06:35 <andythenorth> stefino: coal at both tiles (? tool) and the industry window?
09:09:18 <stefino> it seems that problem is in mail cargo itself. cause all the industry chain gives a coal instead a mail
09:12:43 <stefino> is possible that coal has ID 1 in basic game?
09:14:51 <andythenorth> shouldn't matter if the new cargo is defined correctly :)
09:15:01 * andythenorth wonders why industry grfs are so hard to code
09:15:15 <andythenorth> when I wrote FIRS in nfo 10 years ago, it wasn't this hard
09:26:26 <peter1138> LordAro, hmm, here it's raining now but then dry from 9 til 1. I might risk it.
09:26:47 <andythenorth> 20mph wind here, gusting to what sounds like more
09:26:53 <andythenorth> weather says 'breeze'
09:27:53 <LordAro> peter1138: mm, bit of dampness here too
09:28:14 <andythenorth> apparently long-distance all weather cyclists live longer
09:28:22 <LordAro> beeb weather says 40mph at 11
09:28:57 <peter1138> How can we fix this baseset changes issue? I can't remember what it was.
09:30:27 <LordAro> commit the obs changes, probably :p
09:30:58 <andythenorth> smells all wrong
09:31:14 <peter1138> Who needs to do that?
09:31:26 <peter1138> Hmm, it's web translator commits isn't it?
09:31:58 <andythenorth> so we're committing compiled assets into the repo?
09:32:08 * andythenorth doesn't really understand the issue
09:32:37 <peter1138> Otherwise you need grfcodec to compile. Not sure that's a huge issue?
09:32:57 <andythenorth> and we don't run the compile on the translator commits?
09:33:13 <andythenorth> so the translator commits shouldn't be going straight in?
09:33:20 <peter1138> Although, not sure why it needs grfcodec to build the baseset metadata, mind you.
09:33:20 <andythenorth> they should be in a branch, which is then compiled and merged?
09:33:28 <LordAro> does it need a full compile? i thought obs was a text file
09:35:08 <peter1138> Maybe parts of Makefile.grf.in needs to be split off
09:35:25 <peter1138> I don't know how this works with MSVC either.
09:36:22 <andythenorth> yeah so no_sound.obs has metadata strings, which are presumably translated
09:37:01 <andythenorth> why are we still carrying orig_dos_de.obg
09:37:11 <andythenorth> do we care what was on the German Transport Tycoon CD?
09:37:15 <andythenorth> with one broken sprite?
09:38:21 <andythenorth> presumably that was needed to support people with a specific TTD CD
09:38:26 <peter1138> If we can, why not?
09:38:41 <peter1138> The fact we added stuff to support it suggests it mattered at some point.
09:38:59 <peter1138> And everything is there, so why remove it?
09:39:15 <andythenorth> I was curious, but it makes sense to keep
09:40:22 <andythenorth> ok so openttd.grf *isn't* recompiled
09:40:45 <LordAro> i was gonna say, baseset strings aren't part of the grf anyway, right?
09:40:58 <andythenorth> something recompiled my openttd.grf yesterday
09:41:00 <peter1138> 08:35 <@peter1138> Maybe parts of Makefile.grf.in needs to be split off
09:41:26 <andythenorth> compiles today didn't rebuild openttd.grf so yair
09:41:48 <andythenorth> Makefile.grf.in then?
09:43:45 <peter1138> Yeah could just be in there.
09:45:47 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnwb
09:46:48 <LordAro> shouls be pretty easy to pull out that bit of the makefile (& add it to the eints call)
09:47:26 <LordAro> well, possibly just add it to the eints call, given it'd have to configure otherwise
09:51:09 <peter1138> Oh, there's an issue. When you do git checkout, to replace those baseset files with stock, it uses the *current* time, not the file's committed time.
09:51:26 <peter1138> So after that, make detects that those files are newer and doesn't rebuild anyway.
09:52:37 <peter1138> Someone's written a python script that updates timestamps to the last committed version of a file.
09:53:42 <peter1138> Mind you, we then don't need to store the *.ob* files in the repo, so make clean should remove them.
09:57:53 <andythenorth> specifically the comments
10:06:16 <peter1138> Hmm, I can't make the obg files without grfcodec.
10:06:34 <peter1138> Wait, no, I deleted too much in bin/baseset.
10:09:06 <peter1138> Not going to PR it yet, need to figure out what happens else where.
10:10:49 <LordAro> all: $(OBT_FILES)\nifdef GRFCODEC\nall: <grf files>\nendif might be slightly nicer
10:12:02 <peter1138> You can have multiple all: definitions?
10:12:17 <LordAro> as long as the rule only adds dependencies
10:35:05 *** synchris has joined #openttd
10:45:07 *** chomwitt has joined #openttd
10:50:49 <DorpsGek_II> [OpenTTD/OpenTTD] Alberth289346 commented on pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnro
10:54:46 <nielsm> messing with IDAfree again to reverse amusic.com this time, luckily much of the sturcure is identical to mmusic.com
10:55:04 <nielsm> but I'm continually amazed at how bad the C compiler they used back then was
10:55:34 <nielsm> it makes jumps to jumps as a matter of course, within a single function
10:55:57 <nielsm> and it's not just short/long jump trampolines as far as I can tell
10:58:16 <nielsm> now this may be on the programmers, but it also looks like some things that should be local variables are just randomly placed in uninitialized globals instead
11:02:38 *** Progman has joined #openttd
11:06:12 <nielsm> and memory alignment? what's that
11:06:51 <nielsm> there's a long sequence of words and dwords at odd addresses in the global data
11:11:03 <TrueBrain> you are having fun I see :D
11:12:15 <nielsm> a weird string, never used
11:18:14 <TrueBrain> seems obvious. It is not a 5
11:18:39 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7055: Fix #6408: Improve wording of the dragging signal distance tooltips (juanjo) https://git.io/fhnr9
11:19:13 <TrueBrain> right, okay, so nightlies are blocked till my PR is reviewed. Bit annoying that I cannot apply a pipeline on another branch without going through very odd hoops
11:20:24 <TrueBrain> after that it is figuring out how compatible DigitalOcean really is with an S3 bucket :D
11:21:14 <TrueBrain> and this is working too:
11:21:14 <TrueBrain> [deployer] 2019/01/13 10:20:46 INFO Sentry initialized with release='0.0.0-49-gb1cbfb4' and environment='global'
11:21:29 <TrueBrain> no more hiding crashlogs :D
11:22:44 <LordAro> how long before you get sick of all the custom clients that lie about their versions? :p
11:22:59 <TrueBrain> which custom clients?
11:23:28 <LordAro> doesn't the reddit server use a custom client?
11:23:47 <TrueBrain> where did this conversation took another route, I wonder ..
11:24:17 <TrueBrain> by the lack of PR, I hope it was obvious I did not add Sentry to OpenTTD itself? :)
11:24:38 <LordAro> i've been assuming a thing again, haven't i?
11:24:59 <TrueBrain> I am going to add this to Python applications we are running
11:25:10 <LordAro> that makes much more sense :)
11:25:12 <TrueBrain> in this case, the deployer (hence the [deployer] part)
11:25:48 <TrueBrain> but, an OpenTTD crash log collector is something I would love to look into
11:25:54 <TrueBrain> because .. that would really really help
11:26:13 <TrueBrain> especially if we can automatically analyse stuff, like decode the crash.log etc
11:26:55 <TrueBrain> and we now have the tool to do all that .. but .. first I want to move the current infrastructure I guess :D
11:30:28 <LordAro> TrueBrain: what's filetype_description.txt for?
11:31:37 <TrueBrain> its a file that is now on the mirror, of all places
11:31:43 <TrueBrain> it defines a human readable text per extension
11:31:50 <TrueBrain> I have been wondering if I shouldnt move that to openttd-website
11:32:45 <LordAro> ah, part of manifest.yaml . Wouldn't it be better placed in azure-pipelines/ instead of the root?
11:32:57 <TrueBrain> the root is getting messy in general
11:33:06 <TrueBrain> should I move all the related files in azure-pipelines?
11:33:09 <TrueBrain> (also changelog.sh etc)
11:48:53 <DorpsGek_II> [OpenTTD/website] TrueBrain opened pull request #17: Add: lookup the description based on the extension https://git.io/fhnoU
11:48:59 <TrueBrain> LordAro: well, that was easier than expected :D
11:52:48 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
11:52:56 <TrueBrain> CI will fail on this now, because I changed the location :D
11:55:35 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhnoI
11:56:13 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
11:57:00 <LordAro> i would've been perfectly happy with the "main" ci files being in the root
11:57:23 <LordAro> breaking the CI for all existing PRs (again) will be a pain
11:58:17 <Eddi|zuHause> yeah, we got people barely coping with the last time that happened
12:01:15 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
12:01:26 <TrueBrain> LordAro: it won't break as it did with Jenkins
12:01:31 <TrueBrain> they do, however, need to rebase
12:01:36 <TrueBrain> but they need to do that anyway
12:02:05 <TrueBrain> the Jenkins thingy broke, as we have the CI set to mandatory, and the name changed
12:02:08 <TrueBrain> the name doesn't change now
12:03:08 <TrueBrain> so I guess this only goes wrong if people don't know they should rebase before a new push
12:03:25 <TrueBrain> I can leave behind a temporary redirect, I guess, which we can remove in a week or so
12:03:28 <TrueBrain> not sure that is needed?
12:04:20 <Eddi|zuHause> i don't think "a week" is remotely enough
12:04:41 <TrueBrain> they REALLY should rebase for sure if they push again after a week :P
12:05:03 <Eddi|zuHause> i think you have too high expectations :p
12:05:07 <TrueBrain> master moves too quickly atm to not have that as a requirement tbh :D
12:05:11 <TrueBrain> otherwise things break anyway
12:05:29 <LordAro> it's quite common to have the main CI file in the root anyway, even if you have extra stuff in a subfolder
12:05:44 <TrueBrain> LordAro: yeah, but 3 of those files was getting a bit annoying
12:07:11 <TrueBrain> this reminds me, I still wonder if we should do anything against PRs that do cleanly merge, but should be revalidated
12:07:29 <TrueBrain> for example, that after N commit, a bot automatically requests in the PR for a rebase, and invalidates the CI results
12:08:17 <TrueBrain> also, we should setup a weekly run to check with all kind of configure settings (no network, no zlib, etc)
12:08:32 <Eddi|zuHause> how big should N be?
12:08:51 <TrueBrain> we now allow rebases, because otherwise after every merge you need to rebase all PRs
12:08:53 <Eddi|zuHause> i don't think there can be a meaningful N?
12:08:54 <TrueBrain> that is just too much
12:08:56 <LordAro> N is completely arbitrary, is the issue
12:09:10 <TrueBrain> it can be like .. 7 days
12:09:18 <TrueBrain> but now N is infinite
12:09:22 <TrueBrain> which is the other side of the spectrum
12:09:28 <TrueBrain> and GitHub doesn't have a: rebase this PR, button
12:09:35 <LordAro> if you can do something like "if the file has been modified in master", might be appropriate
12:09:50 <TrueBrain> now I come to think of it, we rebase in the CI
12:09:57 <TrueBrain> so .. it should be enough to just retrigger the CI once in a while
12:10:42 <Eddi|zuHause> and a bot that posts to the PR about rebasing if the once-successful build fails?
12:11:12 <Eddi|zuHause> or a tag needs-rebasing
12:11:26 <TrueBrain> I noticed that people are oblivious to CI results anyway, so a bot to repeat what went wrong might help anyway
12:11:54 <TrueBrain> too often I see people saying that the author needs to fix the broken CI result :D
12:12:22 <Eddi|zuHause> yeah, but that kinda needs better extraction of the actual error
12:12:31 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep opened pull request #7056: Change #6060: Allow drawing dropdown lists with scrollbars above and fix scrolling movement https://git.io/fhnon
12:12:32 <TrueBrain> someone should look into that
12:12:52 <Eddi|zuHause> i wouldn't even know where to begin to look
12:12:58 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
12:13:12 <TrueBrain> there is too much on stderr for linux builds .. so .. that is one
12:13:19 <TrueBrain> and Azure Pipelines only shows 'warnings' for MSVC
12:13:26 <TrueBrain> google might give more details about that
12:16:56 <TrueBrain> LordAro: splitting of the two (Release vs release-stable) really helps in the long-run too :) And the build results are much more clear this way :D
12:21:10 <TrueBrain> talking about not rebasing .... LALALAA :D
12:21:11 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
12:28:42 *** frosch123 has joined #openttd
12:29:00 <TrueBrain> good morning frosch123
12:29:20 <Eddi|zuHause> random thought concerning "real world" savegames in the regression: the closest we have to that is the various title game subnissions?
12:29:44 <LordAro> i imagine only the trunk game has been tested in any significant detail
12:29:45 <TrueBrain> haha, no :D Ever opened a title-game submission? At least, I know of several that are .. euh ..
12:29:48 <TrueBrain> not what you think :D
12:30:30 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhno8
12:30:37 <Eddi|zuHause> uhm, sure, they hide stuff outside the field of view
12:30:43 <LordAro> i imagine there's going to be issues turned up, but screw it
12:30:46 <TrueBrain> yeah .. many were explicitly made for title game :)
12:30:49 <TrueBrain> no "real world" about it :P
12:31:01 <TrueBrain> it doesn't "break" anything :D
12:31:07 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7027: Add: [AzurePipeline] introducing a release pipeline https://git.io/fhGSE
12:31:10 <Eddi|zuHause> but i mean, it could cover more test cases
12:31:32 <TrueBrain> what do you want to test?
12:31:45 <TrueBrain> I once considered keeping an archive of every savegame every reported as bug
12:31:56 <TrueBrain> but we didn't have any infrastructure back then to do anything with that
12:32:09 <LordAro> not like we have anything now :p
12:32:23 <Eddi|zuHause> this is more like fuzzing than having anything specific to test in mind
12:32:31 <TrueBrain> we do have the infrastructure now; well, it needs to be setup
12:32:33 <TrueBrain> but we have the resources :)
12:33:05 <Eddi|zuHause> problem with user-savegames will probably be getting NewGRFs and stuff from non-bananas
12:33:20 <TrueBrain> fair enough; but those are rare as fuck :)
12:33:21 <Eddi|zuHause> also, sorting out patched versions
12:33:44 <TrueBrain> most of that you can automate away
12:33:48 <Eddi|zuHause> but i still think the title games would be a start
12:33:49 <TrueBrain> just dismiss any that are not within parameters
12:34:00 <TrueBrain> either way, go for it :D
12:34:23 <Eddi|zuHause> i'm not the right person for that :p
12:34:42 <TrueBrain> not with that attitude
12:35:17 <TrueBrain> hmm .. right ... IPv6 .. how are we going to do that .. hmm
12:35:37 <Eddi|zuHause> IPv6? that will never catch on...
12:35:46 <Eddi|zuHause> it's been 20 years
12:36:12 <TrueBrain> too bad for you, OpenTTD is 100% ready
12:36:16 <TrueBrain> and I don't want regression
12:36:24 <TrueBrain> but DigitalOcean currently doesn't support IPv6 over the load balancer
12:36:47 <nielsm> hmm, it looks like ADLIB.CAT might actually be containing standard MIDI, and the amusic.com driver just translates that into adlib commands
12:37:34 <Eddi|zuHause> does music from the dos version work yet?
12:37:45 <nielsm> has for several months
12:38:32 <TrueBrain> meh; their CDN is also not IPv6
12:38:37 <TrueBrain> hmm .. that is somewhat of an issue
12:38:54 <Eddi|zuHause> throw it out, rebuild it from scratch
12:39:28 <Eddi|zuHause> what's the fraction of IPv6 traffic nowadays anyway?
12:40:07 <TrueBrain> let me check that for you
12:40:24 *** Progman_ has joined #openttd
12:41:13 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on issue #7043: Disconnected towns during world generation https://git.io/fhnog
12:41:44 <Eddi|zuHause> ok, that's about an order of magnitude higher than what i imagnied :)
12:42:07 <TrueBrain> same for hits on the server
12:42:11 <TrueBrain> that was today .. let me take a bigger sample
12:42:28 <nielsm> okay, adlib.cat is not quite compatible with gm.cat
12:42:49 <nielsm> probably not standard midi, just something reminiscent of it
12:43:21 <TrueBrain> but yeah, I guess this will mean I have to setup an IPv6 -> IPv4 proxy
12:43:28 <TrueBrain> meaning no CDN on IPv6
12:45:28 <TrueBrain> 42.9M hits in 90 days on our webserver; 4.4M of that is via IPv6.
12:45:37 <DorpsGek> TrueBrain: 0.102564102564
12:46:11 <TrueBrain> @calc 4380065 / 42940454
12:46:11 <DorpsGek> TrueBrain: 0.102003229868
12:46:40 <TrueBrain> and that is only the webserver. I am ignoring here things like MasterServer and ContentServer
13:10:07 <peter1138> Hmm, any idea how to create the *.ob? files from a VS project?
13:15:33 <LordAro> similar to how the regression test is done?
13:25:24 <peter1138> Hmm, I guess we can't rely on awk being available.
13:35:47 *** Thedarkb-T60 has joined #openttd
13:37:09 <frosch123> can cmake do such stuff?
13:38:12 <LordAro> it'd be a bit of work, but probably
13:39:05 <frosch123> more work than doing stuff in vbs?
13:39:26 <LordAro> in combination with the rest of the build system? probably
13:45:56 <peter1138> Hmm, that temporary langfiles.tmp is masive.
13:47:15 <LordAro> well, it is every single language file :p
13:59:07 <peter1138> Shame I can't just ignore MSVC :(
13:59:55 <frosch123> how often do you change the basegraphics?
14:00:11 <peter1138> frosch123, I don't.
14:00:40 <LordAro> frosch123: translations change
14:00:49 <peter1138> The issue is when translators update baseset translations, the bin/baseset/* files don't get updated.
14:01:08 <peter1138> I figure if we can just autogenerate those always, instead of including them in the git repo, that would be the best way.
14:01:37 <peter1138> Shouldn't be too hard to make a vbs to generate them.
14:01:47 <frosch123> add a "./configure; make; git add bin" there
14:02:19 <frosch123> if you add vbs. isn't osx the next one to fail?
14:02:41 <peter1138> I'm not touching eints, and including generated files in git is still the wrong thing to do, IMHO.
14:02:56 <peter1138> The fact that they are in git makes building them inconsistent, too.
14:03:15 <LordAro> frosch123: that would work, but it's very ew
14:04:31 <frosch123> maintaining generation scripts for 3 platforms is more ew
14:05:15 <frosch123> you probably lock out all future contributors, and make it a no-touch area
14:05:52 <frosch123> people will need to install grfcodec and stuff
14:05:59 <frosch123> or you implement grfcodec in vbs :p
14:06:54 <frosch123> we generate grf, ob*, squirrel-widgets, squirrel-api, ...
14:09:41 <peter1138> You don't need grfcodec to update the metadata files.
14:10:33 <frosch123> how do you get the md5sum of a container 2 grf?
14:10:57 <peter1138> I'm only touching the baseset metadata generation.
14:11:15 <frosch123> so you keep the obg in the repository, and just add vbs to update them on windows?
14:11:19 <peter1138> Currently that grf is still stored in the repo and I'm not removing it with this change.
14:11:44 <frosch123> obg contains the container2 md5sum of openttd.grf, you won't get that without grfcodec's grfid
14:12:29 <frosch123> maybe it is not container2, but it also eww to disallow that for the future
14:14:32 <frosch123> ok, MAkefile.grf.in currently uses md5sum, not grfid -m
14:16:46 <peter1138> Another problem I found is git doesn't restore timestamps, so make thinks the output is up to date when it's not.
14:17:15 <peter1138> Hence ultimately not including these output files is (an unattainable?) good idea.
14:17:27 *** ChanServ sets mode: +v tokai
14:23:27 <Eddi|zuHause> are you sure you didn't hit a weird config setting with git there?
14:23:43 <Eddi|zuHause> concerning the timestamps
14:24:12 *** Thedarkb-T60 has joined #openttd
14:25:58 <peter1138> No, git always restores files with the current timestamp, not commit timestamp.
14:27:58 <Eddi|zuHause> i find that hard to believe
14:28:24 <frosch123> it's the expected thing
14:28:28 <TrueBrain> time to pick a believe Eddi|zuHause :)
14:28:29 <peter1138> You don't need to believe.
14:28:44 <frosch123> otherwise if you checkout an older reivision, make does nothing
14:29:18 <frosch123> all source files that change need a newer timestamp
14:29:28 <peter1138> This really reinforces "don't commit intermediate files"
14:30:01 <frosch123> Eddi|zuHause: the problem is that both the source and the intermediate file get a new timestamp
14:30:07 <frosch123> and it is luck which one is newer
14:31:06 <frosch123> but it is only a problem if the intermediate files were incorrect in the repo
14:31:32 <frosch123> thus you rely on others (like eints :p) committing correctly
14:31:50 <Eddi|zuHause> which they are for example in bin/baseset
14:32:28 <peter1138> Or if you just happen to git checkout bin...
14:32:39 <peter1138> After making your own changes to the sources.
14:32:59 <peter1138> This is also an issue with the local grf files.
14:33:08 <peter1138> But I'm not suggesting we don't include them, just yet :p
14:33:41 <Eddi|zuHause> so, what's actually the holdup with eints?
14:35:47 <frosch123> noone doing it? i am not aware of any real problems
14:37:05 <frosch123> eints is in some weird intermediate universe between devzone and github
14:40:11 <TrueBrain> orudge: it seems I do not have enough permissions on Azure Pipelines to install addons :(
14:40:48 <Eddi|zuHause> ah... "git update-index --assume-unchanged <file>" should work until eints is fixed :)
14:41:05 <TrueBrain> orudge: and I don't have the password for the main account :P
14:41:18 <TrueBrain> (not stored with the regular crowd, from what I can see)
14:42:52 <TrueBrain> I guess I can reset the password :D
14:47:06 <Alberth> frosch123: hai, no special opinion, openttd seems fine to me
14:49:11 <peter1138> Probably if we rely on it, it should be in OpenTTD.
14:52:07 <Eddi|zuHause> apparently you should rather use --skip-worktree instead of --assume-unchanged
14:53:35 <LordAro> i'd suggest it should be on github, given it's deployed on OTTD infrastructure
14:53:48 <LordAro> you've already agreed that
14:54:07 <Eddi|zuHause> the question was WHICH github
14:55:33 <TrueBrain> Now verifying that encryption works...
14:55:33 <TrueBrain> Not configured. Never mind.
14:56:28 *** andythenorth has joined #openttd
14:57:02 <frosch123> well, there are three versions of eints
14:57:15 <DorpsGek_II> [OpenTTD/OpenTTD] Gabda87 updated pull request #7047: Add #6887: Highlight tiles within local authority of towns https://git.io/fhn44
14:57:26 <frosch123> the generic base version, and then two "forks" with ottd and devzone specific scripts and gui changes
14:57:55 <Eddi|zuHause> "we should have a unified version that covers all use cases"... aka "maka a 4th version"
14:58:24 <LordAro> it's a bit easier when all 3 versions are controlled by the same people :p
14:58:44 <frosch123> so, one repo with 3 branches: master, openttd, devzone
14:58:59 <LordAro> that feels excessively awkward to maintain
14:59:53 <LordAro> first should work out what exactly is different, what *needs* to be different, and probably add that as subfolder in the repo
15:00:58 <DorpsGek_II> [OpenTTD/OpenTTD] Gabda87 commented on pull request #7047: Add #6887: Highlight tiles within local authority of towns https://git.io/fhn6m
15:01:24 <TrueBrain> S3 buckets are weird
15:02:21 <frosch123> reduced user interface for single project, commit scripts and commit message
15:08:12 <DorpsGek_II> [OpenTTD/OpenTTD] Gabda87 commented on pull request #7025: Add #6887: Option to show zone inside local authority boundary of towns https://git.io/fhn6Y
15:09:09 <andythenorth> sounds like deployment-time configuration
15:09:15 <TrueBrain> so, S3 sets Last Modified header always tothe time the file was last touched on the CDN .. which is annoying, as we normally put it on the time of release ..
15:09:16 <andythenorth> but it's been a while since I touched eints
15:21:41 *** sla_ro|master has joined #openttd
15:24:25 * andythenorth pisses around upgrading pip :P
15:24:35 <andythenorth> nothing is ever just 'run this one command' :P
15:24:47 <peter1138> Oh, hmm, that's probably a bug in the RGB patch. 2CC behaves a bit weird :p
15:25:06 * glx wanted to quick fix 7050, but all mingw is broken
15:25:27 <andythenorth> where even is pip :P
15:25:32 <glx> peter1138: issue 7050 is yours btw ;)
15:26:11 <LordAro> glx: is that the duplicate SORT_ASCENDING declaration?
15:26:32 <andythenorth> meh python yak-shaving, that needs a gif :x
15:27:11 <glx> for mingw-w64 yes, but the other mingw is broken because -syd=c++11 defines __STRICT_ANSI__ and removes many windows specific declarations
15:27:38 <LordAro> i've got a small diff to fix the former, anyway
15:27:52 <glx> with -std=gnu++11 it should work
15:29:02 <andythenorth> TrueBrain: what python version do I need for fetch-downloads.py?
15:29:11 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro opened pull request #7057: Fix: A few minor compile warnings under MinGW https://git.io/fhn6E
15:29:30 <andythenorth> I have 3.6 and it's failing
15:29:33 <glx> it fails for strcasecmp, strcasencmp, strdup and many more
15:29:39 <TrueBrain> andythenorth: let me check my magic ball
15:29:52 <TrueBrain> (in other words, be a bit more explicit about 'failing' and I might be able to help :P)
15:30:04 <LordAro> glx: is there any particular advantage to maintain support for oldstyle mingw?
15:30:17 <andythenorth> Successfully installed Coroutine-0.0.2
15:30:26 <glx> probably none as it seems unmaintained anyway
15:30:26 <TrueBrain> import error on dependency .. that is .. euh .. yeah
15:30:40 <TrueBrain> andythenorth: create a new venv, and install everything in 'requirements.txt' in there
15:30:44 <TrueBrain> python -m venv venv
15:30:49 <TrueBrain> venv/bin/pip install -r requirements.txt
15:30:56 <andythenorth> ooh that's a nice command
15:31:07 <TrueBrain> possibly 'python3' instead of 'python'
15:31:10 <TrueBrain> depending on your configuration
15:31:23 <LordAro> there are still some warnings that i can't fix, which annoys me
15:31:24 <andythenorth> yeah so my stupid venv doesn't have pip
15:31:48 <andythenorth> it's right here python-3.6/lib/python3.6/site-packages (python 3.6)
15:31:54 <andythenorth> python -m pip --version
15:32:34 <andythenorth> python -m pip install -r requirements.txt seems to play nicer
15:32:57 <TrueBrain> I never had a situation that after 'python -m venv venv', venv/bin/pip did not exist
15:33:06 <TrueBrain> it should install pip by default
15:33:09 <andythenorth> I delete it and go again
15:33:13 <andythenorth> pip is in the standard lib now
15:33:25 <TrueBrain> but ... Mac might do things weird or something
15:33:33 <andythenorth> it's not a Mac python
15:33:37 <andythenorth> that would never work
15:34:00 <TrueBrain> either way, otherwise run: venv/bin/python -m pip install pip ;)
15:34:37 <andythenorth> i go from scratch and see if I still get import errors
15:34:39 <TrueBrain> btw, for those who care: I renamed OpenTTD-CF to CompileFarm
15:34:47 <andythenorth> one day python might work
15:35:26 <TrueBrain> I use Python 3.6.6 and Python 3.7.something, both work
15:36:04 <andythenorth> meh, it's ignored the frigging env and put them in my python site-packages dir
15:36:07 <andythenorth> I hate this crap
15:36:14 * andythenorth biab after fixing it
15:36:37 <TrueBrain> always use venv/bin/python
15:36:42 <TrueBrain> source venv/bin/activate
15:36:54 <andythenorth> I always use activate
15:36:54 <TrueBrain> in both cases, it should leave your global python installation behind
15:38:22 <andythenorth> I won't give running commentary, we don't need that :(
15:38:41 <TrueBrain> if I can help, let me know :)
15:38:55 <andythenorth> I have a very specific python setup
15:38:57 <andythenorth> you won't like it
15:39:08 <andythenorth> we have to have everything from 2.4 through to 3.6 for work
15:39:13 <TrueBrain> best thing about virtualenvs in Python, that your main setup doesn't really matter :P
15:39:30 <TrueBrain> I myself have a Python installed, from which I bootstrap my other Pythons, as the system Python is foobar :D
15:39:50 <TrueBrain> I even have a venv with both Windows and Linux in it
15:42:23 <TrueBrain> hmm .. 88GB to upload to the CDN
15:42:29 <TrueBrain> how am I going to do this in the most clever way :P
15:43:09 <andythenorth> ok so sometimes 3.6 doesn't include pip
15:43:16 <Eddi|zuHause> because we all know writing to a USB stick is the fastest operation ever
15:43:34 <TrueBrain> andythenorth: python -m virtualenv sometimes helps
15:43:37 <TrueBrain> or virtualenv itself
15:43:41 <TrueBrain> (all 3 different packages)
15:43:53 <glx> Eddi|zuHause: floppy discs ;)
15:44:16 <Eddi|zuHause> glx: good luck attaching 88GB worth of floppydisks to a carrier pidgeon :p
15:44:46 <Eddi|zuHause> @calc 88*2**20/1440
15:44:46 <DorpsGek> Eddi|zuHause: 64079.6444444
15:44:48 <TrueBrain> hmm .. I have to jump two hops .. hmmm
15:45:01 <Eddi|zuHause> it's only 64080 floppies
15:45:15 <andythenorth> ah fuck it I rebuild python
15:45:33 <LordAro> glx: i added lots of nice comments to the PR :)
15:45:44 <LordAro> Samu: not on its own, probably
15:46:41 <Samu> r = 50 - 80 becomes a big number, not negative
15:47:20 <andythenorth> will python build in less than 1 hour? :P
15:47:42 <TrueBrain> do I want to know ....
15:48:11 <Eddi|zuHause> i don't think so
15:50:11 <LordAro> Samu: except that it's + 80 - 80
15:50:12 <andythenorth> python is broken, rebuilding is the only option
15:50:45 <andythenorth> or I could paste some commands from SO
15:50:55 <andythenorth> that curl pip from...who knows where
15:51:00 <andythenorth> and then run them with sudo probably
15:51:35 <glx> negative numbers are big unsigned numbers
15:52:38 <andythenorth> yum python build failed
15:53:16 * andythenorth gives up and uses TrueBrain's zip
15:53:26 <andythenorth> python is absolutely the worst programming environment I ever use
15:53:52 <TrueBrain> that is why I love Docker andythenorth :)
15:53:57 <TrueBrain> it kinda solves that issue
15:54:12 <andythenorth> the punchline is that python is pretty much the only programming environment I use :P
15:55:13 <Samu> no, LordAro, r is a random number from 0 to 255
15:55:19 <Eddi|zuHause> i think it should be safe to make that an int
15:55:30 <Samu> sometimes only r - 80 happens
15:56:06 <Eddi|zuHause> you shouldn't do r-=80 on an uint, if you haven't made sure that it's >=80
15:56:28 <Xaroth> How did you break python, andythenorth?
15:56:29 <TrueBrain> andythenorth: I am guessing, but: docker run --rm -it -v "$(pwd)":/workdir --user $(id -u):$(id -g) python:3.6-alpine /bin/sh
15:56:37 <TrueBrain> that should give you a working Python 3.6
15:56:39 <LordAro> that GB call changes things though, doesn't it?
15:56:49 <TrueBrain> cd /workdir && python fetch-downloads.py
15:56:52 <TrueBrain> should work from there
15:56:59 <TrueBrain> no guarantees; might blow up your machine etc etc
15:57:08 <andythenorth> thanks...but then I invalidate support for my job :D
15:57:17 <andythenorth> we have a specified method for python
15:57:38 <andythenorth> Xaroth: I made a virtualenv
15:57:48 <TrueBrain> andythenorth: that docker command leaves zero traces behind :P
15:57:55 <Eddi|zuHause> it flips a coin, then it skews that coin flip in one or the other direction
15:58:00 <TrueBrain> it only gives you that python till you close the shell :P
15:58:12 <TrueBrain> that is what Dockers are for
15:58:23 <andythenorth> the zip is nice too :P
15:58:28 * andythenorth is looking at the downloads page
16:00:33 <andythenorth> where is the official download for pip? :P
16:00:49 <andythenorth> not some random dude on SO with a fork
16:01:03 <andythenorth> is pypi dead yet?
16:01:55 <Eddi|zuHause> $fvrtpkgmgr install pip
16:02:27 <andythenorth> wrong answer Eddi|zuHause, that gets me outside of support for work :P
16:03:07 <andythenorth> soon we will be off python2.4
16:03:08 <Eddi|zuHause> your work tells you to make things more complicated than necessary?
16:03:13 <andythenorth> we have no choice
16:03:25 <andythenorth> we have 2.4, 2.7, and various 3
16:03:37 <andythenorth> we are close to killing 2.4
16:04:16 <andythenorth> ha ha, I now have a not broken python3.6 env
16:04:22 <Xaroth> 2.4 has been out of support for over a decade O_O
16:05:02 <andythenorth> but if you're on Plone 3 with no migration path, it's a huge upgrade to 2.7
16:05:09 <andythenorth> like 18 months work
16:05:45 <andythenorth> I had to live through it
16:06:48 <andythenorth> IT'S NEARLY OVER
16:07:19 <andythenorth> "OpenTTD-website.andythenorth/env/bin/pip"
16:07:30 <andythenorth> somebody broke how virtualenv works
16:07:43 <Xaroth> but by the time your migration to 2.7 is done, 2.7 will be out of support :P
16:07:44 * andythenorth used the old one
16:08:18 <andythenorth> 2.7 to 3 isn't so bad
16:09:36 <Samu> it was uint r since svn r1
16:11:06 <Samu> im touching code that has never been touched in the last 15 years
16:11:40 <andythenorth> TrueBrain: I'm sure you'll be excited, I fixed python :P
16:12:26 <TrueBrain> I will be if you approve my PR :P
16:12:38 <andythenorth> I switched to 3.7, script works
16:12:49 <andythenorth> I have some crap version of 3.6 where it fails
16:13:24 <Samu> i notice it doesn't support 90 degrees
16:13:32 <Samu> the forbidding of 90 degrees
16:13:55 <andythenorth> forbidding 90 degrees for ships is clown shoes
16:15:27 <Xaroth> Hey, I can't help it I have large feet, andythenorth.
16:15:34 <Xaroth> no need to call them clown shoes!
16:18:03 <Eddi|zuHause> i vote for removing the "disallow 90° for ships", and instead allow continuing on other trackbits on the same tile (effectively allowing 360° turns on tile borders)
16:18:46 <Eddi|zuHause> that should prevent a bunch of "ship is stuck" problems
16:23:27 <peter1138> I have a patch for that
16:26:09 <Eddi|zuHause> PR or it didn't happen
16:27:14 <DorpsGek_II> [OpenTTD/website] andythenorth approved pull request #17: Add: lookup the description based on the extension https://git.io/fhnit
16:28:11 <andythenorth> at some point PRs are going to overhaul issues, in the racing sense of overhaul
16:28:47 <Eddi|zuHause> "we now have more solutions to problems than problems"?
16:29:06 <andythenorth> in some cases yes
16:29:09 <andythenorth> competing approaches :)
16:29:50 <andythenorth> probably ripe now
16:30:10 <peter1138> Hmm, is there still an OPF for ships?
16:30:22 <andythenorth> didn't you delete it?
16:30:54 <peter1138> Hrmm, not sure why it touches yapf_road.
16:32:37 <Eddi|zuHause> i thought we only made yapf the default
16:33:12 <peter1138> Can we not just remove non-Yapf?
16:33:18 <Eddi|zuHause> peter1138: but that patch only adresses the first half of my request, not the "allow turning into a trackbit on the same tile"
16:33:19 <peter1138> Hasn't it proven itself yet?
16:33:39 <Eddi|zuHause> people citing fringe cases...
16:34:02 <peter1138> Hmm, other trackbits that don't actually connect?
16:34:03 <Samu> forbid 90 deg helps pathfinders search less tracks, but may also make it harder to build canals inland
16:34:25 <Eddi|zuHause> assume you come from a track bit to a tile border
16:34:35 <Eddi|zuHause> pathfinder checks the 3 connecting trackbits on the next tile
16:34:54 <Eddi|zuHause> but for ships it might also check the 3 trackbits on this tile (including the one we just came from)
16:35:56 <Eddi|zuHause> peter1138: there were cases where a ship got stuck on a diagonal trackbit, because an industry got built blocking both entrances of that trackbit
16:36:10 <Eddi|zuHause> peter1138: if this reversing could happen, it could choose a different exit of that tile
16:36:21 <Eddi|zuHause> by switching trackbits
16:36:45 <peter1138> Maybe allow that only if there are no connecting track bits on the neighbouring tile.
16:36:58 <frosch123> andythenorth: original pathfinder is gone for trains and rv
16:37:00 <Eddi|zuHause> i don't see a reason for that restriction
16:37:02 <Samu> i made a patch request for that Eddi|zuHause
16:37:05 <frosch123> it's still there for ships
16:37:34 <peter1138> Because allowing all bits always is more expensive, as Samu pointed out.
16:38:23 <Eddi|zuHause> maybe allow all bits only on the origin tile of the pathfinder request, assume deeper in the pathfinder there are always better paths
16:38:43 <andythenorth> when is NPF deprecated? o_O
16:38:55 <Samu> when an industry builds on water, this check is checked
16:39:15 <Samu> basically it makes ship behave like they're 3 tiles wide
16:39:23 <Samu> for construction purposes
16:40:48 <frosch123> it's needed everytime yapf desyncs
16:41:10 *** Thedarkb-T60 has joined #openttd
16:41:18 *** Wormnest has joined #openttd
16:45:40 <DorpsGek_II> [OpenTTD/website] TrueBrain merged pull request #17: Add: lookup the description based on the extension https://git.io/fhnoU
16:45:55 <LordAro> peter1138: i would not be opposed to removing !YAPF
16:46:27 <LordAro> iirc, OPF got left in for ships because YAPF was horribly slow at ships, but that got fixed
16:46:45 <frosch123> LordAro: problem with removing yapf is that only michi understands it
16:46:47 <Eddi|zuHause> now it's just slow, not horrible? :p
16:46:53 <frosch123> npf otoh is really easy to maintain
16:47:23 <frosch123> so, if you want to go for a single pf, i would prefer to remove !NPF :p
16:47:24 <TrueBrain> we should collect statistics of what people are using :P
16:47:32 <andythenorth> don't rivers use one of them to generate? o_O
16:47:55 <frosch123> TrueBrain: people obviously use the default, that's easy
16:48:09 <andythenorth> frosch123: which of the built-in economies is the stupid one?
16:48:15 * andythenorth wonders about removing that
16:48:41 <frosch123> how about focusing on broken stuff, instead of trying to break working stuff?
16:48:49 <andythenorth> I just like fire
16:49:10 <LordAro> frosch123: but YAPF is so much better?
16:49:18 <andythenorth> fire is so very cleansing
16:49:23 <frosch123> LordAro: not for maintainers :p
16:49:41 <peter1138> yapf isn't that bad, when you realised it's based on templates.
16:49:50 * andythenorth deletes some Horse stuff
16:50:05 <andythenorth> not even committed, no way to get it back now ::)
16:50:33 <Samu> im gonna try make forbid 90 deg work on OPF
16:50:57 <andythenorth> reasons aren't reasons
16:51:07 <Samu> unless you really are going to remove it for ships
16:51:26 <LordAro> Samu: ultimately, original is ...old
16:51:33 <LordAro> if it's broken, it shouldn't be fixed
16:51:44 <LordAro> depending on how broken it is
16:52:01 <LordAro> or rather, you shouldn't be trying to make it better
16:52:06 <DorpsGek_II> [OpenTTD/OpenTTD] gregcarlin commented on pull request #7003: Feature #6918: Add option to adjust font size separately from GUI size. https://git.io/fhnig
16:58:12 *** Flipp3rrr has joined #openttd
17:01:26 <Samu> i'm afraid it's gonna be too long if I do it
17:04:21 *** snail_UES_ has joined #openttd
17:05:34 <LordAro> Samu: not knowing much about it at all, how does "The tile where the ship actually "enters" the depot is the north tile. Ship depots are also valid water tracks for ships from any company, unlike the other depot types, so the pathfinder has to do a "blind" search for a tile that matches these conditions." sound?
17:06:44 <LordAro> comments don't have to be short
17:06:49 <Samu> and doesn't explain why I'm doing the change to be this way
17:07:09 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPv
17:07:12 <LordAro> comments should explain *why* the code is doing what it is
17:07:42 <LordAro> it shouldn't explain what it is doing (that's in the code) nor why code was changed (that's for commit messages)
17:08:13 <Samu> pathfinders before never did the blind search
17:08:28 <Samu> the depot tile was given to them directly
17:08:34 <Samu> meh, it's becoming long again
17:09:07 <glx> oh make depend fails /bin/sh: D:/developpement/GitHub/glx22/OpenTTD/objs/release/depend.exe: Argument list too long
17:09:57 <glx> that explains why modifying a header didn't trigger recompile
17:10:28 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7003: Feature #6918: Add option to adjust font size separately from GUI size. https://git.io/fhnPI
17:11:17 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPL
17:12:29 <Eddi|zuHause> so, what's the policy on multi-commit PRs? squash them? make sure every revision builds individually? (CI doesn't seem to do that)
17:12:41 <peter1138> Gah, bloody Phantoms. They're so lethal :-(
17:13:11 <LordAro> Eddi|zuHause: rebase by preference, unless there's clearly no need for the extra commits
17:13:33 <LordAro> i wouldn't say there's any great need to test every revision
17:13:40 <Eddi|zuHause> LordAro: i don't understand what that means
17:14:03 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro updated pull request #7057: Fix: A few minor compile warnings under MinGW https://git.io/fhn6E
17:14:04 <LordAro> Eddi|zuHause: which bit? :p
17:14:26 <Eddi|zuHause> "<LordAro> Eddi|zuHause: rebase by preference, unless there's clearly no need for the extra commits" <-- this
17:14:49 <Eddi|zuHause> LordAro: the original was 5 patches, i made 5 separate commits out of those
17:15:56 <Eddi|zuHause> LordAro: but when i tried to compile just the first patch, it didn't compile
17:16:04 <peter1138> I tend to do merge commits on my own repos, means my local branches can be 'safely' deleted.
17:16:15 <peter1138> But I guess that would get messy.
17:16:31 <LordAro> Eddi|zuHause: trickier when it's not your patch, i guess
17:16:50 <LordAro> i'd definitely recommend trying to make it compile
17:17:03 <Eddi|zuHause> LordAro: the question is, relocate stuff between patches until the queue works at every stage, or just squash everything
17:17:33 <LordAro> that's ultimately up to you, but individual commits are easier to review, and it is quite a large number of changes
17:18:11 <Eddi|zuHause> there's also a savegame bump in there, and changes related to savegames in separate patches of the queue
17:19:30 <LordAro> oh, this isn't #7051, is it?
17:21:01 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 commented on pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPc
17:23:07 <glx> LordAro: custom OTTDxxx in os_abstraction.h are no longer needed
17:23:33 <glx> I commented the whole else block and it compiled fine
17:23:39 <LordAro> ah, well that makes it a lot easier
17:23:51 <Samu> It's not worth caching the result with NPF_FLAG_IS_TARGET
17:24:02 <Samu> what is this FLAG, where is it
17:24:19 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN dismissed a review for pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPL
17:24:20 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN updated pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPv
17:24:59 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN updated pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPv
17:25:15 <Samu> if I can speed up NPF ship depot checking, then it's worth investigating what this NPF_FLAG_IS_TARGET is all about
17:29:47 <peter1138> Hmm, one image per sprite, or one sprite sheet per logical grouping?
17:30:25 <peter1138> At least, in the GUI stuff.
17:32:39 <Samu> isn't fibonnacy queue faster for pathfinders?
17:33:51 <glx> even if one per sprite is better for vcs
17:36:28 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick updated pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhZzm
17:40:35 <andythenorth> logical grouping peter1138
17:40:45 <andythenorth> I looked yesterday, there are a few obvious groups
17:40:52 <andythenorth> e.g. there's some autorail and depot stuff
17:41:05 <andythenorth> a load of groups icons
17:41:54 <andythenorth> looks like there's icons for window chrome
17:42:58 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro updated pull request #7057: Fix: A few minor compile warnings under MinGW https://git.io/fhn6E
17:45:15 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas opened issue #7059: Town name language choice affects number of towns / world population https://git.io/fhnPP
17:50:12 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPH
17:50:24 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7058: Fix #7050: Missing guards around network chat message functions for compiling with networking disabled https://git.io/fhnPv
17:51:10 * andythenorth drives from London to Newcastle
17:52:15 <LordAro> pretty sure #7059 is a duplicate of #6640
17:53:30 <nnyby> it's similar, but #6640 doesn't mention how town name language affects the number of towns as well as world size
17:54:06 <LordAro> indeed, the suggested patch isn't a complete fix by any means
17:54:18 <LordAro> but the issue is ultimately the same
17:54:40 <LordAro> i'll probably close the older issue anyway
18:22:12 *** sla_ro|master has joined #openttd
18:28:36 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth approved pull request #7056: Change #6060: Allow drawing dropdown lists with scrollbars above and fix scrolling movement https://git.io/fhnXS
18:29:02 <andythenorth> hmm, I didn't actually review the code style there
18:29:53 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth dismissed a review for pull request #7056: Change #6060: Allow drawing dropdown lists with scrollbars above and fix scrolling movement https://git.io/fhnXS
18:30:13 <andythenorth> ^ anyone want to review that ^
18:30:47 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6078: Orders: it does not jump order by remaining lifetime https://git.io/fhnX9
18:31:36 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6155: Vehicles are sometimes incorrectly shown as refittable https://git.io/fhnXH
18:35:43 <Samu> just found a bug with NPF
18:36:34 <Samu> manual sending a ship to enter depot right after leaving it, makes it reverse 180 degrees
18:36:44 <Samu> doesn't happen with opf or yapf
18:42:24 <Samu> ah i see why, reverse penalty cost is 0
18:44:45 <peter1138> Bah, the clicky wheel button on my mouse no longer makes the wheel clicky :(
18:45:54 <Samu> hmm, no, actually the problem is deeper
18:47:56 <glx> andythenorth: can you add a . between find and -type at squirrel_export.sh:29 and try it on your system ?
18:51:05 <Samu> npf pathfinding to a tile where the vehicle is at the moment will make it return INVALID_TRACKDIR. For ships that means it will reverse
18:52:04 <Samu> it should walk a 180 turn and come back to the tile
18:52:06 <glx> yes but you can test it in master too
18:52:35 <Samu> pathfinder pretty much refused to pathfid
18:52:37 <peter1138> K, this is classy, I decided to light up my minecart track with... jack-o-lanterns :p
18:52:55 <andythenorth> peter1138: that's the default approach
18:53:12 <andythenorth> I had to build one like that once
18:53:18 <glx> and torches on only one side in the caves
18:53:24 <peter1138> Is it? I've never done that.
18:53:33 <andythenorth> I guess you're not 5
18:53:37 <peter1138> Yeah, always done that. I want to avoid torches down the minecart tunnels though.
18:53:50 <andythenorth> illuminate them with TNT
18:54:09 <andythenorth> I could never take minecraft seriously, I just used to set fire to everything
18:54:14 <glx> use creepers if you want to destroy stuff ;)
18:55:37 <andythenorth> glx squirrel_export.sh runs now for me, it's pretty angry about L54 tho
18:55:38 <andythenorth> ./../squirrel_export.sh: line 54: [: -eq: unary operator expected
18:56:56 *** wodencafe has joined #openttd
18:57:04 <peter1138> Creepers tend to just destroy me.
18:57:18 <peter1138> I've put on the full diamond armour now, although the enchantments are crap because, well, I died.
18:57:41 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened issue #7060: NPF refuses to pathfind when the destination tile is the same as where the vehicle currently is https://git.io/fhn1C
18:58:09 <peter1138> Why should it pathfind in that case?
18:58:30 <Samu> because there's tracks ahead of the ship
18:58:41 <andythenorth> you don't need to pathfind if you're on the tile
18:58:46 <andythenorth> that makes sense no?
18:59:02 <glx> andythenorth: add " around the backquoted stuff before -eq and replace -eq with =
18:59:26 <Samu> ship is in tile and moving at direction NE
18:59:38 <glx> the -eq is not present in master btw
19:00:08 <Samu> now it's ordered to go to the same tile, when the ship reached the end of the tile, it does a 180 turn, now moving SW
19:00:09 <glx> looks like a change I shouldn't have made ;)
19:00:36 <Samu> the other pathfinders will make the ship walk a few tiles outside to make the turn
19:00:52 <andythenorth> glx: seems pretty happy after those changes
19:01:06 <glx> good, I'll add them to the PR
19:01:18 <andythenorth> the ship is told to reverse no?
19:01:21 <Samu> the bad behaviour is the reversal when the path isn't block
19:01:23 <andythenorth> NE is the opposite of SW?
19:01:33 * andythenorth probably lacks context but nvm
19:01:34 <Samu> meh i fail at explaining
19:01:47 <Samu> let me draw something, brb
19:01:50 <glx> I guess you got only "generating ..." lines and no "updated" nor "deleted"
19:05:14 <Samu> the other pathfinders will walk that path when told to go to the same place the ship is
19:05:15 <Xaroth> Looks a bit like a train
19:05:37 <Samu> npf however will make the ship reverse at the end of the tile, doesn't walk
19:07:08 <andythenorth> why do they need to walk around?
19:07:14 <andythenorth> just reverse the ship
19:07:25 <Samu> because that's "good behaviour"
19:08:33 <andythenorth> seems a bit pointless polishing a dead pathfinder :P
19:08:54 <Samu> it's kinda the same as if it was a train or a road vehicle, reversing when there's tracks ahead looks ugly
19:18:52 <DorpsGek_II> [OpenTTD/OpenTTD] JGRennison commented on issue #6470: Link graph job thread-join schedule changes https://git.io/fhn1y
19:26:40 <DorpsGek_II> [OpenTTD/OpenTTD] JGRennison commented on issue #6472: Cargodist: various link graph performance improvements https://git.io/fhnMJ
19:26:46 <peter1138> Ships can actually turn on the spot, trains can't.
19:27:00 <peter1138> Also, I have a patch that makes ships turn slowly of instantly.
19:30:39 <Samu> at the very least, the behaviour is inconsistent with the other pf's
19:32:25 <andythenorth> peter1138: delete ships!
19:32:40 <andythenorth> peter1138: make ships great again!
19:32:54 *** gelignite has joined #openttd
19:36:42 <Samu> let me see what happens to ships going to 1 buoy only
19:38:35 <Samu> npf is the one with the reversal behaviour
19:55:45 <Samu> cost += _settings_game.pf.npf.npf_buoy_penalty; // A small penalty for going over buoys
19:56:05 <Samu> why the cost for going over buoys is a penalty?
19:56:49 <LordAro> i imagine that the buoy operator would be fairly annoyed if you ran over it with a ship :p
19:57:37 <Samu> it could even be a bonus
19:57:51 <Samu> but ppl might abuse buoys
19:57:54 <Eddi|zuHause> manned buoys... great idea
19:58:39 <LordAro> Samu: gameplaywise, i do agree with you, buoys shouldn't affect pathfinding at all
19:58:43 <LordAro> you could make that a PR :)
20:01:03 <Samu> must investigate what's the current cost first
20:04:24 *** Thedarkb-T60 has joined #openttd
20:05:17 <Samu> there had to be a reason someone made it a penatly
20:05:42 <andythenorth> stop ships over-running bouys?
20:10:28 <Samu> - Add: [NPF] Ships get a penalty for going over buoys now, so they will try to go around.
20:10:48 <Samu> for "realistic" purposes
20:11:24 <Eddi|zuHause> andythenorth: that only works if you move buoys to tile corners
20:12:11 <Eddi|zuHause> or enforce that buoys use a 3x3 area
20:13:51 <Samu> guess there's no need for a PR
20:14:09 <Samu> there is a game setting for it, can be edited to 0
20:17:55 <andythenorth> we stayed up late
20:18:04 <andythenorth> then I woke up early
20:18:55 <andythenorth> TrueBrain: you should leave encouraging comments
20:18:57 <andythenorth> on the older PRs
20:19:07 <TrueBrain> COME ON, ALMOST THERE
20:19:18 <andythenorth> peter1138 has the oldest
20:19:35 <andythenorth> Samu has the most
20:19:43 <andythenorth> and also the most esoteric
20:22:39 <nielsm> hm well adlib.cpp compiles without warnings or errors, but it's still massively incomplete
20:23:04 <peter1138> Do we need adlib support?
20:23:26 <LordAro> peter1138: do we need cargodist?
20:23:35 <peter1138> Yes, I use cargodist.
20:23:36 <Eddi|zuHause> do we "need" anything?
20:23:46 <nielsm> (exactly the result I expected)
20:23:53 <peter1138> But if you have original_dos or original_tto support, is original_adlib necessary?>
20:23:59 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on issue #5881: Allow separate compilation (objdir != srcdir) https://git.io/fhnMp
20:24:10 <LordAro> random question, is STR_COLOUR_RANDOM used anywhere?
20:24:26 <peter1138> Yes, it's used when setting your default company colour preference.
20:24:31 <DorpsGek_II> [OpenTTD/OpenTTD] JGRennison commented on pull request #6917: Fix iconv and clang version detection on OSX https://git.io/fhnDv
20:24:32 <Eddi|zuHause> that would be my question as well, is the adlib.cat music really any different from the gm.cat music?
20:24:36 <peter1138> It was recently added.
20:24:57 <nielsm> the adlib soundtrack is different in instrumentation, but right now I'm just doing this because I've wanted to do it for a long time
20:25:07 <nielsm> I agree it's largely superfluous
20:26:02 <Eddi|zuHause> i should maybe look for my copy of TTO
20:26:15 <nielsm> the adlib.cat music format looks much like the format in gm.cat, but it's not quite the same
20:26:45 <nielsm> for one thing, each song header contains a list of extra instrument definitions in addition to the 17 built-in programs
20:27:11 <nielsm> so the adlib format can technically program the synthesizer rather flexibly
20:28:00 <nielsm> (you could argue that with sysex you can also program a good hardware synth rather flexibly via standard midi, except most people won't have those hardware synths)
20:29:55 <Eddi|zuHause> nope, can only find old TTDP installations
20:30:23 <Eddi|zuHause> and i don't have any FDD drives available
20:30:30 <Eddi|zuHause> to go through my old disks
20:30:45 <Eddi|zuHause> which probably aren't readable anymore anyway
20:31:54 <TrueBrain> andythenorth: possibly good to keep a few of those "good first issue", despite them being stale
20:32:02 <TrueBrain> some really are trivial, and a good way to get started with OpenTTD
20:32:08 <TrueBrain> not sure .. just spitting out a random idea :P
20:32:29 <nielsm> (yes that demo with only toyland playable and just one or two maps)
20:32:43 <nielsm> (but it does have the full music collection)
20:32:54 <Eddi|zuHause> i had a TTO demo once
20:33:18 <Eddi|zuHause> which i maybe still have
20:33:26 <Eddi|zuHause> but i'm not going to look for it
20:33:35 <peter1138> I think I still have a few mid-90s compilation CDs from magazines.
20:35:19 <andythenorth> TrueBrain: yeah I applied stale pretty liberally, but 'good first issue' is good for new contributors
20:38:44 <andythenorth> it escalates from 'scroll fix' to 'tile height refactor'
20:38:48 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #6917: Fix iconv and clang version detection on OSX https://git.io/fhnDG
20:40:23 *** Thedarkb1-T60 has joined #openttd
20:42:39 <Eddi|zuHause> "warning: refname '6583' is ambiguous." ... uhm...
20:42:57 <andythenorth> conflates with a PR? :P
20:43:12 <andythenorth> nice number, 5999 is pleasing shape
20:43:25 <Eddi|zuHause> i named the branch like that, but it is probably the beginning of an existing hash
20:45:00 <Eddi|zuHause> no idea why it didn't properly apply
20:46:51 <Eddi|zuHause> ah, i think comment changed
20:47:57 <Samu> damnit, debug mode is too slow for 5000 ships
20:49:58 <Eddi|zuHause> they do that sometimes
20:51:22 <peter1138> Strange diff format :p
20:51:28 <andythenorth> don't really understand.rej
20:51:35 <andythenorth> seems to have not information
20:52:45 <Eddi|zuHause> try "patch --merge"
20:53:10 <Eddi|zuHause> puts the rejected/conflicted sections in the code with <<<<< ===== >>>> markers
20:53:21 <andythenorth> seems my patch doesn't have --merge
20:53:54 <andythenorth> looks like just formatting, and add one line?
20:55:26 <andythenorth> meh, those aren't the failed hunks
20:57:33 <Eddi|zuHause> great, another comment conflict
20:58:00 *** Thedarkb2-T60 has joined #openttd
21:00:54 <Samu> experimenting no random track choice for opf, see how much it affects performance
21:03:56 <Samu> strange, i expected better performance, i get the opposite
21:04:10 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z opened pull request #7061: Fix #6583: Rework Tileheight handling (patch by adf88) https://git.io/fhnDK
21:06:44 <Samu> wow fast forward is ultra mega fast now
21:08:22 <Samu> 9999 frames per sec, i like
21:08:46 <peter1138> You can speed it up further if you really want.
21:08:46 <nielsm> it loads the index of a song now, at least
21:08:53 <nielsm> then on to actually playing it
21:09:27 <Samu> let me test fast forward on 4k maps
21:09:44 <peter1138> Will probably not be much different from before.
21:09:58 <nielsm> peter1138 I'm going to squash things before proposing a PR :P
21:10:30 <peter1138> Ok, you added the OPL emulator. I'm interested now :p
21:10:45 <peter1138> That's a lot of code for a frivolous feature, mind you.
21:11:24 <peter1138> Damn, my Linux VM has no audio :(
21:11:28 <Eddi|zuHause> why is windows build taking so much longer than linux?
21:13:02 <peter1138> 4kx4k is limited by tick processing time, not graphics.
21:14:39 <Eddi|zuHause> skimming through that patch queue i wonder if the first "remove unused functions" patch really should be the first
21:25:55 <glx> Eddi|zuHause: optimisations
21:26:10 <glx> and it's the linking phase that takes time
21:26:19 <glx> for debug builds it's the oposite
21:26:20 <nielsm> ugh the hardest problem in making this adlib decoder/player is probably to correctly synchronize the music ticks with pcm sample output
21:26:51 <glx> oh and file accesses are slow as hell on windows too
21:27:52 <andythenorth> hmm someone crashed into a van in ETS, and it wasn't me
21:28:14 <glx> nielsm: realtime timer in the music thread
21:28:39 <andythenorth> all the cars in ETS 1, I have either owned, or lived with someone who has
21:28:47 <andythenorth> probably means I have to buy a truck next :P
21:29:27 <andythenorth> oof why is the in-game cursor sprite so laggy in ETS1?
21:29:36 <nielsm> glx, I'm not sure I can thread this
21:29:52 <nielsm> since I'm bound to the game's mixer callback to render the music
21:30:07 <nielsm> gives me a buffer and expects me to fill it completely
21:30:29 <nielsm> and that buffer can be quite long, I think often 8192 samples on windows
21:30:58 <peter1138> Isn't the mixer callback already threaded?
21:31:08 <glx> but the buffer filling has to be threaded anyway
21:34:43 <nielsm> the OPL emulator fills the PCM buffer it's given with sound generated for whatever register settings it has at the time of the buffer fill call, i.e. time only advances for the emulator when it's told to generate some samples
21:35:44 <nielsm> so I have to read some music commands, fill in OPL registers, determine delay until next batch of commands, generate samples for that delay, and repeat
21:35:50 <glx> sound thread already calls mixer functions directly
21:36:27 <nielsm> and through that keep track of remainders and rounding due to one timer running at 44.1 kHz and another running at 66.7 Hz
21:37:07 <peter1138> You can use floats there though :p
21:38:45 <glx> yes floats are only banned from game logic
21:43:18 <glx> oh I'm stupid, relation between mixer and audio is the opposite
21:43:33 <glx> audio drivers read the mixer
22:02:25 <glx> oh the mixer never used the callback feature
22:02:28 *** Thedarkb1-T60 has joined #openttd
22:07:14 <glx> I think it should be possible for the OPL emulator to fill an internal buffer, and copy its content on callback
22:11:39 <glx> when you register the callback you get the sample rate, but you won't know the buffer size to fill before the actual callback call
22:13:41 <glx> as it's created by the audio driver, then passed to the mixer, then to the callback if any
22:27:41 *** andythenorth is now known as Guest595
22:27:41 *** andythenorth has joined #openttd
22:50:59 <Samu> messing around with opf, making it less random
22:51:38 <Samu> wondering what's the difference between DistanceManhattan and DistanceMaxPlusManhattan in terms of results
22:57:24 <DorpsGek_II> [OpenTTD/website] LordAro closed issue #11: Make the copyright in the bottom automatically bump years https://git.io/fhOMR
22:57:35 <andythenorth> new contributor?
22:58:14 <DorpsGek_II> [OpenTTD/website] andythenorth commented on pull request #18: Change copyright date to use site.time https://git.io/fhnSg
22:58:36 <andythenorth> drive by contributors TrueBrain ^ :)
23:01:31 <LordAro> trickle down eco^H^H^Hcontributions!
23:04:40 <andythenorth> the article is worth scanning, maybe not on Sunday night, but eh
23:04:59 <andythenorth> proposes that drive-through contributions are a sign of healthy project
23:05:14 <TrueBrain> and that it was easy to access
23:05:17 <LordAro> TrueBrain: given staging's copyright link still says 2018, i assume it doesn't get automatically updated
23:05:52 <TrueBrain> check pipeline to see where it is at
23:06:03 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7056: Change #6060: Allow drawing dropdown lists with scrollbars above and fix scrolling movement https://git.io/fhnSX
23:06:47 <TrueBrain> shows 2019 for me now
23:07:52 <andythenorth> good ottd weekend
23:07:53 *** andythenorth has left #openttd
23:08:11 <TrueBrain> it is not instantly :p
23:08:50 <TrueBrain> this contribution alone made it worth the effort :)
23:10:46 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7025: Add #6887: Option to show zone inside local authority boundary of towns https://git.io/fhnSy
23:10:56 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on issue #7060: NPF refuses to pathfind when the destination tile is the same as where the vehicle currently is https://git.io/fhnSS
23:12:40 <Samu> found another reversal bug, now with OPF
23:13:04 <Samu> but maybe it's caused by me, maybe not a bug
23:15:03 <Samu> why can't i change branches when i have stuff edited
23:15:16 <Samu> i dont want to commit these edits yet
23:16:17 <LordAro> `git stash pop` to reapply once you've switched
23:21:06 <Samu> looks like it also happens on upstream/master
23:21:22 <Samu> savegame is incompatible, can't test in 1.8.0 yet, brb
23:25:53 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro requested changes for pull request #7049: Fix #6599: Can still click on buy button in vehicle selection window even if no vehicle is selected https://git.io/fhn9I
23:27:33 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7046: Fix #5978: Update some stats and lists caused by group hierarchy. (3298 and Juanjo) https://git.io/fhn9L
23:29:19 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7046: Fix #5978: Update some stats and lists caused by group hierarchy. (3298 and Juanjo) https://git.io/fhnlj
23:29:24 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro closed issue #5978: Autoreplace window isn't updated after group hierarchy changes https://git.io/fhn9q
23:41:54 *** AKTheKnight has joined #openttd
continue to next day ⏵
