IRC logs for #openttd on OFTC at 2018-04-30
⏴ go to previous day
00:04:39 *** Supercheese has joined #openttd
00:30:33 *** Stimrol has joined #openttd
00:33:59 *** Thedarkb-X40 has joined #openttd
00:40:10 *** Stimrol has joined #openttd
01:39:34 *** Eddi|zuHause has joined #openttd
01:40:31 <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
04:22:47 *** DuckDuckPenis has joined #openttd
04:29:56 *** DuckDuckPenis has left #openttd
04:34:48 *** Supercheese has joined #openttd
04:47:03 *** Laedek_ has joined #openttd
04:51:07 *** muffindrake4 has joined #openttd
05:12:44 *** HerzogDeXtEr1 has joined #openttd
06:06:55 <Eddi|zuHause> ah you gotta love windows update... "something didn't work"
06:07:00 <Eddi|zuHause> with no information what
06:43:39 <Eddi|zuHause> also fun is how the windows versions apparently use different versioning system
06:44:57 <Eddi|zuHause> "the running version is 14393, the running version is 16299"... what happened to versions 1607,1709 etc?
07:10:19 <Eddi|zuHause> nope... update still stopping halfway through
07:47:47 *** sla_ro|master has joined #openttd
07:58:12 *** andythenorth has joined #openttd
08:05:16 *** smoke_fumus has joined #openttd
08:22:56 *** som89__ has joined #openttd
09:25:25 <Arveen> the 5 digit numbers is the build number, the 4 digit one the release name
09:41:13 <Eddi|zuHause> sure... but why? i have no clue which is which...
09:41:29 <Eddi|zuHause> and surely 99.9999% of users neither
09:41:54 <Eddi|zuHause> but whatever, apparently consistency is too much to ask
10:02:54 <LordAro> marketing reasons mostly, aiui
10:12:32 *** Cybertinus has left #openttd
10:17:49 *** sla_ro|master2 has joined #openttd
10:32:31 <peter1138> So direct harfbuzz?
10:32:50 <peter1138> Until that gets dropped :p
10:33:34 <peter1138> libreoffice uses harfbuzz, to answer one of the questions in that bug.
10:49:09 <Arveen> 1803 / 17134.5 is not listed yet, but will be added soon I guess
10:55:48 <blathijs> peter1138: The problem is that, AFAIU, harfbuzz only offers layouting, not wordwrapping
10:56:51 <blathijs> peter1138: I saw that post before, it also says that harfbuzz doesn't do wrapping (and looking at harfbuzz's site, I don't think that changed since then)
10:57:35 <blathijs> I can imagine Libreoffice and others implement their own word-wrapping on top of harfbuzz, since the layouting is probably the tricker part of the two problems (though looking at the ParagraphLayout code, wordwrapping isn't entirely trivial either)
11:16:09 <Eddi|zuHause> Arveen: the existence of that page is completely pointless, since it's not accessible from the "everyone should understand this" updater program
11:19:36 <Arveen> but that's why we have "something didn't work" blue screens - so everybody understands what's going on :D
11:25:06 *** andythenorth has joined #openttd
11:32:07 <andythenorth> Pikka: so 1200hp small engine then :P
11:32:30 <andythenorth> I've brought back the Little Bear
11:32:34 <andythenorth> it was kind of iconic for the set
11:33:21 <andythenorth> I tried an 8/8 edition but it looks a bit weird
11:46:46 <andythenorth> ok subject to play-testing, horse engines all done
11:46:51 <andythenorth> I said that twice before :P
11:49:51 *** sla_ro|master2 has quit IRC
12:07:47 *** FLHerne has joined #openttd
12:53:02 <peter1138> Hmm, should I make a 2x/4x gui set?
12:53:19 <peter1138> Or are the doubled-up pixels "cute"?
12:53:37 <andythenorth> the original base set scales up fine at 2c
12:53:47 <peter1138> I always get annoyed by the alignment errors :(
13:12:07 *** andythenorth has left #openttd
13:26:26 <peter1138> Bollocks, I edited the wrong text field :p
13:36:33 <FLHerne> I thought there was an EZ gui already
13:37:18 <peter1138> Unless it's been updated, it's from before GUI zoom was implemented, and so the sizes are wrong.
13:38:07 <FLHerne> Yes, just discovered that
13:38:41 <FLHerne> Presumably you could steal the actual sprites
14:06:14 <peter1138> They may be in OpenGFX. But I don't use that.
14:35:57 *** MPAAA3X6 has joined #openttd
14:45:09 *** LorenzoDV has joined #openttd
14:46:32 *** andythenorth has joined #openttd
14:46:39 <LorenzoDV> do you know if nightlies are being built from the new git repo?
14:47:56 <andythenorth> oh I am behind :P
14:48:09 <andythenorth> I might be spreading FUD
14:48:25 <andythenorth> I thought binaries were coming from existing CF, and they'll be building last rev from svn
14:48:53 <andythenorth> it's wavey hands though, I am not canonical :P
14:49:25 <Wolf01> I'm too much used to compile it by myself
14:50:39 <peter1138> They're not being built yet.
14:51:25 <peter1138> CI is making builds just to make sure it compiles, they're not published though.
14:51:51 <LorenzoDV> ok... I'll finally have to set up my built environment then!
14:54:29 <LorenzoDV> @peter1138: yes, in fact I just found the Jenkins server dashboard doing CI, but no artifacts
14:54:44 *** snail_UES_ has joined #openttd
15:00:24 <blathijs> peter1138: LordAro: TrueBrain: Just compiled OpenTTD 1.8.0 against ICU60 and icu-le-hb (latter two were prepared by the Debian ICU maintainer in experimental, so no changes to OpenTTD required) and that seems to work. However, AFAUI this approach is fragile, potentially buggy and not really maintained, so it might still be worthwile to look for an alternative.
15:00:54 <peter1138> We've got a ton of ICU-related crashes in the bug reports too.
15:01:45 <blathijs> Would we still need ICU if we switch to Harfbuzz + some word-wrapping implementation? Or does Harfbuzz do everything we need from ICU?
15:05:46 <LordAro> ICU is also used for unicode sorting
15:07:12 <peter1138> So collation. Hmm./
15:08:56 <LordAro> as far as i know, there have been no issues with that and it's not deprecated :p
15:14:15 <blathijs> Yeah, and string splitting (in words and characters) it seems (curiously guarded by WITH_ICU_SORT)
15:15:59 <blathijs> It also links against LEFontInstance, but that is probably layout-related and provided by Harfbuzz as well
15:32:03 <blathijs> But it would probably be fine to use both ICU and Harfbuzz side by side (I believe Harfbuzz even has some optional ICU integration, though I can't find anything specific about this just now0
16:03:55 *** Alberth has joined #openttd
16:03:55 *** ChanServ sets mode: +o Alberth
16:22:49 <MPAAA3X6> are you THE Charles?
16:26:23 *** supermop_work has joined #openttd
16:35:12 *** Lejving__ has joined #openttd
16:48:47 *** Wormnest has joined #openttd
16:58:39 <andythenorth> more retconning :P
16:58:47 <andythenorth> retcon was my word of 2017 :P
17:05:50 <Alberth> rediscovering your own hidden vehicles? :)
17:06:08 <Pikka> does AI brakevan, andythenorth?
17:06:20 <andythenorth> Pikka: do you mean 'should it', or 'empirically'?
17:08:02 <andythenorth> Alberth: one of the reasons Horse exists is because pikka wouldn't add that engine to UKRS 2 :)
17:08:13 <andythenorth> but I'd entirely forgotten about it
17:08:16 <Pikka> loco + tender + brakevan is a lot for a TL3 though
17:08:41 <Alberth> cargo transport is optional :p
17:08:46 <andythenorth> how do you detect brakevan?
17:09:42 <Alberth> and an add-on was a too simple solution, I guess, andy? :)
17:10:30 <Pikka> by assuming any wagon with 0 capacity is a brakevan, naturally :)
17:10:48 <andythenorth> Alberth: something something 'limited roster of trains'
17:11:05 <andythenorth> can't remember whose idea Horse was :)
17:11:10 <Alberth> before the pixels exploded thus
17:11:39 <Alberth> I think it was a good idea anyway :)
17:13:43 <andythenorth> "DanMacK Tue Sep 03, 2013: Ideas for KISS train Set"
17:16:23 <andythenorth> I think we tried 8 trains originally
17:16:46 <andythenorth> supermop_work: want to play "name that train"?
17:17:00 <andythenorth> I need a name that's like Janus, the two headed god
17:17:08 <andythenorth> that's an actual UK engine
17:17:13 <andythenorth> and it can be mythical
17:17:24 <andythenorth> I also have Chimera and Phoenix in UK Horse now
17:17:49 <andythenorth> oh I need names for GWR-style railcars too
17:18:02 <andythenorth> 'Banana' and "Bananagram" are placeholder :P
17:18:55 *** synchris has joined #openttd
17:22:51 *** ChanServ sets mode: +v tokai
17:23:47 <supermop_work> how many heads does cerberus have?
17:55:18 *** andythenorth has joined #openttd
18:19:12 *** frosch123 has joined #openttd
18:22:29 <Wolf01> <andythenorth> by accident <- how do you add a diesel engine by accident?
18:23:10 <Wolf01> Meh... he's blinking again
18:23:34 <frosch123> now even the last one should figure out, that andy is the author of xkcd
18:29:30 <Wolf01> Now I understand why the characters of xkcd sometimes resemble some of this chat people :D
18:30:35 *** Progman has joined #openttd
18:33:55 *** andythenorth has joined #openttd
18:38:30 *** chomwitt has joined #openttd
18:38:30 <andythenorth> supermop_work: "Big Keith"? o_O
18:38:35 <andythenorth> it replaces "Little Bear"
18:49:10 <frosch123> LordAro: 6770 does not add any warning for gcc6 or clang3.8. do you have other compilers for testing?
18:49:38 *** andythenorth is now known as Guest1340
18:49:39 *** andythenorth has joined #openttd
18:49:42 <LordAro> gcc7 is the one that introduced all the warnings
18:49:46 <LordAro> i'll test when i get home
19:01:37 *** sla_ro|master has joined #openttd
19:17:43 *** andythenorth has joined #openttd
19:26:05 <andythenorth> Wolf01: it's worse than that
19:26:26 <andythenorth> they missed Apple Python
19:26:37 <andythenorth> python.buildout (which actually solves all this btw)
19:26:54 <andythenorth> another setuptools
19:27:01 * andythenorth spamming the channel, oops
19:27:32 <andythenorth> they missed this "export PATH=$PY:$PY26:$PY24:$PY27:$PY32:$PY33:$PY34:$PY35:$PY36:$PYPY3:$PATH"
19:28:46 <andythenorth> it's all basically fine
19:28:54 <andythenorth> AS LONG AS YOU NEVER PASTE THINGS FROM STACK OVERFLOW INTO SHELL
19:29:01 <andythenorth> which is a good rule for life anyway
19:34:25 <Eddi|zuHause> frosch123: even more dirty than trying to make sense of a windows installation that a completely computer-illiterate person used for 2 years?
19:34:56 <andythenorth> just reinstall every year or so
19:35:03 <glx> first step: remove all browser toolbars
19:35:15 <andythenorth> things are worse than I though
19:35:29 <frosch123> Eddi|zuHause: that involves personal contact, so it depends on the person
19:36:31 <Eddi|zuHause> frosch123: it's a close relative, that you don't have any feud with
19:36:59 <frosch123> i probably do not even know them
19:37:27 <Eddi|zuHause> i was speaking hypothetically
19:37:34 <Eddi|zuHause> i hope you do know some of your relatives :p
19:38:26 <andythenorth> your close relatives aren't allowed to use Windows?
19:46:08 <TrueBrain> right, nightly releases .... I have to change a Jenkins plugin for that ...NOT looking forward to that :(
19:46:11 *** supermop_work has joined #openttd
19:47:53 <TrueBrain> *pokes andythenorth about #6753* :D
19:51:04 <frosch123> (# in first column being the CERT recommendation to work around bugs of 30 year old compilers)
19:51:24 <frosch123> (# indented being easier to explain)
19:51:45 <TrueBrain> no opinion here; dont give a hoot :)
19:51:51 <nielsm> indentation after # is what I'm most used to seeing
19:54:24 <glx> means openttd source code itself will be rejected
19:55:31 <frosch123> tb's beos pr was rejected because it indented after # :)
19:56:44 <TrueBrain> and you dont write pre-commit-hooks for what is in source; but what is added to it :P
19:58:48 <andythenorth> all the trains for 1950-2050 should be visible in the buy menu at once
19:59:00 <andythenorth> the 2 I just added don't fit, I should delete them
19:59:28 <andythenorth> 'designing for specific screen sizes' :P
20:08:54 <frosch123> ok, even google code style says "do not indent preprocessor" stuff
20:09:08 <frosch123> now i only need someone to press "approve" :p
20:09:45 <TrueBrain> did NOT read the patch :P
20:13:42 <TrueBrain> guess that means I have to do a riddle again too ...
20:17:21 *** ericnoan has joined #openttd
20:21:31 * andythenorth reads how to checkout PRs .... again
20:21:59 <frosch123> fetchpr = !sh -c 'git fetch $1 pull/$2/head:pr$2' -
20:23:33 <andythenorth> 6753 is tricky for me, but I'll test
20:23:35 <TrueBrain> updated commit-checker on CI; but we should also make a PR to update it .. couldnt be bothered to follow the right way :P
20:23:42 <frosch123> LordAro: everything with args needs sh afaik
20:23:46 <andythenorth> I turn off all the trackpad gesture shit on the mac
20:23:49 <andythenorth> except rotate :P
20:23:49 <TrueBrain> andythenorth: the main question is IF that is expected :P
20:23:58 <andythenorth> I'd have to turn it on and test
20:24:10 <andythenorth> I'd have to use some Apple apps too
20:24:12 <TrueBrain> none of us use OSX enough to know how other apps do that :P
20:24:18 <TrueBrain> you are awesome, tnx :)
20:24:26 <andythenorth> I don't use any of that stuff either
20:24:41 <frosch123> TrueBrain: we have a second pr, i couldn't be botherer to make separate updates :p
20:24:51 <TrueBrain> we will wait for that :P
20:25:08 <andythenorth> also my compile is really slow, can we fix that, thx :P
20:25:39 <andythenorth> it's more than 2 mins
20:26:32 <andythenorth> wow scroll map on scrollwheel would be really nice
20:27:17 <andythenorth> why did I turn it off?
20:27:26 <TrueBrain> because you are weird
20:27:51 <andythenorth> oic it was broken
20:35:26 <andythenorth> TB notification incoming :P
20:35:52 <TrueBrain> still not the answer to the question we asked :D
20:35:58 <TrueBrain> IS THIS WHAT OTHER OSX APPS DO :P
20:36:09 <TrueBrain> and tnx; that was pretty quick tbh :)
20:36:23 <andythenorth> just reword this in your head
20:36:24 <andythenorth> "pinch-to-zoom confirmed now works per macOS conventions"
20:36:32 <andythenorth> until it answers your question :P
20:36:44 <TrueBrain> totally read over the last part of that sentence
20:36:47 <andythenorth> I hate 'macOS' name it's ugly as fuck
20:37:03 <andythenorth> it even violates there brand guidelines
20:37:08 <andythenorth> it's "Mac" with capital M
20:38:07 <andythenorth> TrueBrain: child A tested the CF binaries at the weekend btw
20:38:20 <TrueBrain> tell me you didnt do A/B testing :P
20:39:41 <andythenorth> interesting concept though
20:39:51 <TrueBrain> but what was his/her ruling?
20:39:54 <andythenorth> I have crash logs, but they're on a mac somewhere else
20:40:01 <andythenorth> I'll dig them out
20:40:19 <andythenorth> and multiple hard crashes
20:40:31 <andythenorth> but always the same 'const char tried to fill' shit
20:40:34 <andythenorth> which is ICU no?
20:40:36 <TrueBrain> was it more stable, ignoring the crashes?
20:40:48 <andythenorth> yes, ignoring the crashes it was stable :P
20:40:59 <andythenorth> oh and AIs OOMing
20:41:06 <andythenorth> the cursor artefacts are gone
20:41:15 <TrueBrain> that is good news at least
20:41:19 <TrueBrain> AIs OOMing is just silly
20:41:57 <andythenorth> I have a cursor bug in on macOS 10.12 which wasn't in the binaries child A was playing on 10.13
20:42:04 <andythenorth> but is for the new CF binaries
20:42:14 <andythenorth> cursor locks to left-side of screen
20:42:21 <andythenorth> making game unplayable
20:42:31 <andythenorth> I keep trying to video it
20:42:37 <andythenorth> but as soon as I click, it resolves
20:42:48 <TrueBrain> sounds like you solved it :P
20:42:53 <TrueBrain> you enable an auto-clicker!
20:42:58 <andythenorth> sometimes I wonder if we should just drop the game cursor and use the OS
20:43:03 <andythenorth> there's a FS request about it
20:43:25 <andythenorth> well it was in Flyspray
20:43:32 <andythenorth> and now is probably in GH somewhere
20:43:38 <andythenorth> but we probly closed it
20:43:38 <TrueBrain> and closed, most likely :P
20:44:12 <andythenorth> macOS hates custom cursors
20:44:19 <andythenorth> and seems some WIndows does too
20:45:25 <TrueBrain> Subject: LAST & FINAL WARNING NOTICE!!!
20:45:32 <TrueBrain> so many things wrong
20:45:42 <TrueBrain> last and final ... how is that different? This is your last, but not your final!
20:45:45 <TrueBrain> and how many !!!!!!!
20:46:08 <TrueBrain> This settlement is about unpaid bequest and last warning notice
20:46:10 <TrueBrain> to you regarding the death of my Late client Edward and his
20:46:11 <TrueBrain> family while on holiday at New Orleans Louisiana USA in the
20:46:13 <TrueBrain> August 2005 Atlantic hurricane Katrina disaster.
20:46:14 <TrueBrain> what the hell is written here ...
20:47:55 <andythenorth> buy AWS time with it
20:48:01 <TrueBrain> I should tell him that
20:48:07 <TrueBrain> reply: pay my AWS with that money!
20:48:39 <andythenorth> tell him: file a PR
20:50:41 *** ToffeeYogurtPots has joined #openttd
20:56:24 <andythenorth> I think that's enough of those
21:10:32 *** ToBeFree has joined #openttd
21:27:06 <peter1138> Well, we should use OS cursor where it's possible to update the image for it?
21:27:13 * V453000 finally finished his awesome drag&drop tool for spritesheets and 8bpp converting :)
21:27:18 <peter1138> Even webpages can do that these days :S
21:29:06 <nielsm> andythenorth you don't have any "wardrobe"(?) style shunters?
21:30:37 <peter1138> Yay, got my compile warnings being tracked properly now. Had to write a script to munge the log file as it didn't have full paths, so Jenkins couldn't show the source files.
21:32:32 <andythenorth> the system cursor appears and disappears frequently on OS X build
21:32:45 <andythenorth> nielsm: the challenge of little engines like that is....why? :)
21:32:59 <andythenorth> even if I give them OP stats, they have no gameplay role
21:34:00 <andythenorth> with real-ish stats
21:34:07 <peter1138> Then when someone(TM) implements shunting, you'll be there for it.
21:34:17 <andythenorth> it's easy to do anyway
21:34:35 <nielsm> a tiny short railway extending a dock's range to a slightly inland industry
21:34:45 <nielsm> and wanting as short as possible trains
21:34:54 <andythenorth> I should add rope-haulage :P
21:35:20 <nielsm> have anyone made horse-drawn trains?
21:35:36 <andythenorth> horse trams though
21:35:41 <supermop_work> i find horses are not that good at drawing trains
21:35:53 <nielsm> they can't hold the pencil?
21:35:55 <supermop_work> they can't hold the pen in their hooves
21:36:07 <andythenorth> supermop_work: you should be kbanned :P
21:38:56 <andythenorth> mornington crescent also
21:40:42 <andythenorth> is that a livery colour?
21:41:28 <peter1138> I should finish some patches :p
21:43:10 <Pikka> andythenorth, where are the superchinooks? :D
21:43:53 <andythenorth> Pikka: considering them :P
21:49:35 <V453000> OH COME ON IT'S JUST A TRAIN
21:59:49 <andythenorth> we still have an issue because he didn't get the Arctic ship last time round
21:59:57 <andythenorth> 'collection is not complete' :x
22:01:41 <frosch123> how many hours do they have to test iron horse until they get something like that?
22:06:15 <andythenorth> shall I train them to respond to feature requests also?
22:06:34 <frosch123> respond how? closing?
22:06:52 <andythenorth> or verifying they're valid also
22:08:54 <peter1138> andythenorth, why is it still "LEGO CITY"
22:09:15 <peter1138> All those arctic cities...
22:13:24 <peter1138> Some Jungle theme...
22:14:10 <nielsm> it feels silly waiting for all the linux and osx builds for a patch that only affects win32 :(
22:14:50 <andythenorth> anyone want to write a web-based HP-per-ton calculator for ottd train physics? :P
22:14:54 <andythenorth> or I just test in game
22:15:23 <frosch123> you would just put the formula into python
22:15:28 <frosch123> so no need for web stuff
22:17:26 <andythenorth> so I could just set the speed a 250t train should go?
22:17:36 <andythenorth> and then the HP is calculated automatically for the newgrf
22:18:32 <frosch123> i guess you could set uphill speed, and time to accelerator to max speed on flat track
22:19:06 <andythenorth> might be over-engineering it :)
22:19:36 <andythenorth> I could just OP all trains
22:19:45 <andythenorth> then there's no need to worry
22:19:51 <andythenorth> "HP: more than enough"
22:22:38 <glx> <nielsm> it feels silly waiting for all the linux and osx builds for a patch that only affects win32 <-- and windows builds are not even tested ;)
22:23:52 *** beno is now known as Thedarkb
22:28:15 <andythenorth> michi_cc: what fix am I testing? o_O
22:28:39 <michi_cc> Mouse cursor getting stuck.
22:29:15 <andythenorth> it's hard to trigger, but I am running a game right now anyway
22:30:08 <andythenorth> I should raise a bug report
22:30:12 <andythenorth> so you can do a PR against it :P
22:30:25 <michi_cc> Right click scrolling, mouse out of the window to the left, mouse back in was very triggery for me.
22:31:05 *** rocky1138 has joined #openttd
22:33:56 *** FLHerne has joined #openttd
22:36:21 <andythenorth> michi_cc: can't trigger it so far, but really brief test
22:42:05 <michi_cc> LordAro: Which makes $compiler become 'Apple'
22:43:04 <LordAro> wonder if it'd just be easier to see what defines are present
22:44:38 <michi_cc> Is there a pure sh grep equivalent? Then you could check if the first output line contains clang *somewhere*.
22:48:09 <LordAro> it's also a bit moot if TrueBrain is redoing the build system with cmake
22:53:46 <nielsm> hmm yeah integrating dosbox's opl2 emulation in ottd and decoding the dos version fm music through it is probably fine
22:54:18 <nielsm> I'm comparing dosbox playing the music to old recordings from an actual hardware fm synth and they sound pretty much identical
22:54:27 <nielsm> good job, dosbox people (and ken silverman)
22:57:06 <glx> I guess it's a line between those airports
22:59:04 <nielsm> and it's put "full load" orders on both stops
22:59:15 <nielsm> so it takes forever to load on the left of them
23:06:06 <nielsm> this is also the version of the game where you servicing an industry is an almost-certain death sentence for the industry
23:06:32 <nielsm> if you play with "disasters" on, as the Medium and Hard difficulties have
23:14:26 <TrueBrain> michi_cc: why on earth does OSX not have __BYTE_ORDER ... its POSIX since 2001? Ugh
23:15:03 <TrueBrain> (and why did it work on the CI .. :P)
23:15:28 <TrueBrain> with the SDK of OSX
23:15:42 <TrueBrain> it cannot use the headers of the host system
23:15:51 <TrueBrain> bit surprised, but also a potential issue
23:15:56 <michi_cc> It might come from a compiler header and not a SDK header.
23:16:11 <TrueBrain> owh well .. chances of hitting a BE system are ..... well .. low
23:16:34 <glx> could be a compiler define too
23:16:42 <michi_cc> Don't expect an Apple clang to come with real clang headers :)
23:17:02 <TrueBrain> because why would they ...
23:17:18 <glx> it even doesnt come with the original compiler name
23:17:24 <TrueBrain> also just a prefix with DARWIN .. so silly ...
23:17:35 <TrueBrain> but I am happy you fixed it michi_cc :)
23:18:18 <michi_cc> Bit hard to find any other compiler errors if you get about 8 lines of warnings for every single source file :)
23:18:39 <TrueBrain> I was hoping the CI represented the real thing a bit more, but meh
23:18:44 <TrueBrain> also didnt check for warnings tbh
23:18:47 *** supermop_work has joined #openttd
23:18:50 <glx> indeed __BYTE_ORDER seems to be a compiler macro, so not related to SDK
23:19:20 <glx> makes sense for it to be absent from apple "custom" compiler ;)
23:20:00 <TrueBrain> the clang is compiled against something in the SDK
23:20:06 <TrueBrain> without an SDK I cannot make the clang work for OSX
23:20:12 <TrueBrain> so I was hoping for a bit more realistic
23:20:17 <TrueBrain> guess I was hoping wrong
23:20:21 <TrueBrain> but what more issues does it present ..
23:23:18 <FLHerne> Why is train weight "0..1279 tons" ?
23:24:14 <TrueBrain> the question you should be asking .... why not?!
23:24:16 <FLHerne> In the NFO spec it's defined as "prop.24*256+prop.16", where both are one byte, so why doesn't it go up to 65k?
23:24:38 <FLHerne> (not that I particularly /want/ a train to weigh 65 kilotons
23:25:35 <michi_cc> LordAro: I was a tiny bit evil :p
23:25:36 <FLHerne> (and yes, that definitely says it's in tons and not silly multiplied units)
23:29:35 <FLHerne> Hm, and in the NML source it's just `two_byte_property(0x16, 0x24, {'unit_type': 'weight'}, {'unit_type': 'weight'})`
23:29:47 <FLHerne> So where does the 1279 come from?
23:30:45 <FLHerne> Prop24 would have to be limited to only two bits, which doesn't make much sense
23:30:45 <TrueBrain> LordAro: why did you need to know if it was clang or gcc again? Don't defines do?
23:31:02 <peter1138> So we have this option in settings for "scrollwheel scrolls map"
23:31:12 <peter1138> It's only implementedd for OSX.
23:31:32 <LordAro> TrueBrain: different set of flags
23:31:45 <TrueBrain> ah .. thought they were compatible
23:31:48 <TrueBrain> okay, that is annoying
23:33:22 <TrueBrain> just surprised clang is not compatible with gcc parameters :P
23:34:01 <andythenorth> peter1138: is that a good or bad thing? o-O
23:34:18 <peter1138> andythenorth, it's code I didn't really know about, and can't test anyway.
23:36:25 <LordAro> TrueBrain: it's basically compatible, but different versions with different warning flags and so on
23:36:58 <TrueBrain> the reason I guess cmake just compiles a few applications to detect all kinds of shit
23:37:07 <TrueBrain> (they dont run it; they inspect the resulting binary)
23:42:08 <peter1138> Damn, accidentally did a rebuild :(
23:42:30 <TrueBrain> guess the CI queues on a monday are also in the order of hours :P
23:42:48 <TrueBrain> main issue to fix that, is that the GitHub plugin Jenkins has, forces a Jenkinsfil in the repo
23:42:53 <TrueBrain> which is a very insecure way ...
23:45:13 <TrueBrain> where did your parser fail?
23:47:15 <FLHerne> Is it me, or does the NML example_train grf not build?
23:47:27 <LordAro> TrueBrain: why are the CI queues busy?
23:47:36 <FLHerne> It uses shorten_vehicle, which according to the changelog hasn't existed since 0.3
23:47:38 <TrueBrain> because people keep making PRs? :P
23:47:40 <LordAro> or, why is them being busy relevant to us?
23:47:56 <TrueBrain> because our PRs take for ever to validate?
23:48:12 <TrueBrain> well, maybe not relevant to you, as you don't have a PR :D
23:48:47 <LordAro> oh, they take that long?
23:49:01 <LordAro> that seems excessive, even for running 3 separate builds
23:49:53 <TrueBrain> yeah, had to wait 30 minutes for a slot
23:49:59 <TrueBrain> build takes ~20 minutes
23:50:09 <LordAro> ah ok, that's less bad
23:50:21 <TrueBrain> just too many people push stuff :D
23:50:22 <LordAro> needs more builders :p
23:50:37 <TrueBrain> which is not possible without altering the GitHub plugin of sorts
23:50:53 <TrueBrain> or maybe I should just do the VCS commands myself .. hmm
23:51:09 <LordAro> why is it not possible?
23:51:31 <TrueBrain> it needs to load the Jenkinsfile from the VCS (for what-ever stupid reason); I cannot force one
23:51:38 <TrueBrain> so I want to run that in a Docker for obvious reasons :D
23:51:56 <TrueBrain> and then you get in a bit of a silly shit, that multiple jobs can starve the CPU
23:52:16 <TrueBrain> I rather have that I can force a Jenkinsfile, and not have the initial job run in a Docker
23:52:22 <TrueBrain> than the normal Jenkins max executor can do its thing
23:52:24 <nielsm> hmm random useless measurement: a full release build for win32 on my desktop takes about 3 minutes
23:52:25 <TrueBrain> solving the whole mess
23:52:29 <LordAro> why would you need to force a jenkins file?
23:52:42 <TrueBrain> would you trust a random Jenkinsfile from a random PR?
23:52:54 <LordAro> well no, but that's why docker, right?
23:52:58 <peter1138> Yeah, that's a bit of a design flaw :S
23:53:13 <TrueBrain> that is why I have to load the initial Jenkinsfile in a docker
23:53:14 <LordAro> could have some sort of "needs validation if jenkinsfile modified" thing
23:53:26 <TrueBrain> that would be possible .. but then I need a feedback system
23:53:32 <TrueBrain> so that means a bot on GitHub
23:53:37 <TrueBrain> which is going to happen sooner o rlater anyway ..
23:53:46 <TrueBrain> but same issue, different solution .. still shitty :P
23:54:06 <TrueBrain> yeah, I want to be able to do something like: @DorpsGek: create release
23:54:16 <TrueBrain> and he gives 1 hour later URLs to download release binaries
23:54:19 <TrueBrain> as some PRs I would really like to test
23:55:13 <TrueBrain> also I noticed that it would be nice if you can say: @DorpsGek: autoclose 1 month
23:55:22 <TrueBrain> and that after a month of no activity, he closes the issue :P
23:55:27 <TrueBrain> as fuck, I am not going to be able to track that :D
23:56:14 <TrueBrain> and DorpsGek needs replacement too .... so yeah ..
23:56:17 <TrueBrain> either way, first the CI
23:57:20 <TrueBrain> just shitty those plugins force you in 1 way
23:57:25 <TrueBrain> when ever was 1 way the correct way for any CI
23:57:35 <TrueBrain> their solution is: only trust contributors
23:57:44 <TrueBrain> like .... that is shitty :P
continue to next day ⏵