IRC logs for #openttd on OFTC at 2020-07-19
⏴ go to previous day
01:10:36 *** snail_UES_ has joined #openttd
02:44:02 *** snail_UES_ has joined #openttd
06:33:15 *** WormnestAndroid has quit IRC
06:34:58 *** WormnestAndroid has joined #openttd
07:15:53 *** tamilguy has joined #openttd
07:16:58 *** tamilguy has joined #openttd
07:52:38 *** DorpsGek_III has joined #openttd
07:52:38 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz dismissed a review for pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJmhB
07:52:41 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz updated pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJTYj
07:57:39 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz commented on pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJnFV
07:57:49 <tamilguy> I have a sent a mail to translator week ago for Tamil translation. I haven't got reply back. Should I wait or send another one?
07:57:49 <tamilguy> I like to contribute to Translations.
08:20:26 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz commented on pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJnbB
08:22:02 *** WormnestAndroid has quit IRC
08:22:15 *** WormnestAndroid has joined #openttd
08:55:46 *** Smedles has joined #openttd
09:06:30 *** Smedles_ has joined #openttd
09:11:57 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz commented on pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJnNS
09:18:19 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz updated pull request #8260: Codechange: Spell 'Viewport' consistently https://git.io/JJTYj
09:34:30 *** HerzogDeXtEr has joined #openttd
09:42:29 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz updated pull request #8217: Port GUI rendering improvements from JGRPP https://git.io/Jf9yI
10:04:57 *** Progman has joined #openttd
10:06:48 <Eddi|zuHause> a 14 year gravedig, is that a record?
13:15:15 <_dp_> amount of nice tree grfs is pretty much 0 :(
13:15:59 <_dp_> well, japanese are ok I guess but too japanese :p
13:33:18 *** frosch123 has joined #openttd
13:41:17 <Flygon> We need a good Australian Tree Pack.
14:11:17 <b_jonas> I've seen photos of Australia. they have normal trees like the ones we have here too, not only eucalyptus.
14:11:50 <frosch123> didn't they all burn down?
14:30:50 <b_jonas> ok wait, it's not *quite* normal, I found the reference:
14:31:10 <b_jonas> it says Australia mostly has evergreen trees
14:31:13 <b_jonas> so no, not quite normal trees
14:31:19 <b_jonas> but there are trees other than eucalyptus
14:31:29 <b_jonas> so yes, there might need to be a tree pack
14:32:49 <_dp_> one of the things I learned from geoguesser is that australia is big and diverse :p
14:33:12 <_dp_> # and has shitty stretview quality
14:35:00 <b_jonas> yeah, I know. let's say this is about the east coast.
14:49:25 <_dp_> is it possible to have objects autoremovable by anyone?
14:50:03 <_dp_> I mean make newgrf object that acts that way
15:02:31 <frosch123> OBJ_FLAG_ANYTHING_REMOVE <- that way it behaves like fields
15:04:33 <_dp_> frosch123, you sure? looks like that only works for same company
15:05:10 <_dp_> or OWNER_NONE with magic bulldozer
15:05:14 <frosch123> oh, yes. i thought you wanted to place them via mapgen
15:07:25 <_dp_> frosch123, oh, I'm reading it wrong, it does work with OWNER_NONE
16:11:08 *** ChanServ sets mode: +v tokai
16:11:48 *** jottyfan has joined #openttd
16:43:04 <supermop_Home_> most non-eucalyptus trees in Australia are introduced / non-native
16:44:44 <supermop_Home_> certainly outside of towns you see mostly all eucalyptus species - but like lemurs in Madagascar or birds in NZ, there are lots of different types tht have evolved to fill various niches
16:45:23 <supermop_Home_> its not just all the same quintessential gum tree everywhere
17:00:52 *** snail_UES_ has joined #openttd
18:09:37 <_dp_> hm, I managed to make an object that crashes the game xD
18:25:00 <nielsm> yeah it should at most generate a grf loader error and fail the entire file
18:39:07 <michi_cc> glx: CMake MSVC regression project fails for me with "(1 more lines were expected than found)". As far as I can determine by manually running the equivalent command, the output is correct. Any idea how to debug that?
18:41:21 <michi_cc> Also, even if I remember it working for me before, the lang files are not generated in a valid position in MSVC. Working dir (with AI compat files) is <foo>/bin, exe dir is <foo>/build/Debug, and langs are <foo>/build/lang, which is not in OTTD search path.
18:45:24 <michi_cc> Both problems do not occur if using Open directory/Ninja to build.
18:45:42 <frosch123> _dp_: when building over it?
18:46:20 <_dp_> frosch123, crash you mean? no, when building object
18:46:30 <frosch123> ok, that's new then
18:46:30 <_dp_> something with 0 size selection
18:47:09 <frosch123> there is a lot of special code for overbuilding multi-tile objects, so i was worried you found another case where it does not work
18:50:53 <glx> michi_cc: IIRC compile farm uses MSVC project too and it works fine
18:51:14 <michi_cc> But still 2017, right?
18:52:06 <michi_cc> I've looked into the CMake files, but it it all directly piped into some CMake variables, and I have no idea how to e.g. get debug output from them.
18:52:10 <glx> yes Visual Studio 15 2017
19:03:07 <michi_cc> glx: Problem is that REGRESSION_RESULT does not end with a newline (which I guess ERROR_STRIP_TRAILING_WHITESPACE ensures), while REGRESSION_EXPECTED does end with a newline.
19:03:59 <michi_cc> Well, replace newline with ";". Maybe some of the regexes before mess up.
19:21:49 *** jottyfan has joined #openttd
19:27:28 <_dp_> any good way to make smooth transition from object tiles to ground tiles?
19:27:38 <_dp_> without making 256*19 sprites
19:41:43 <nielsm> just generate your way out of it, 5000 sprites is not a problem if you just have to write a loop!
19:47:23 <frosch123> _dp_: draw the ground tile first, then add your partial-transparent layers
19:47:37 <frosch123> i.e. use the ground tile of the baseset, not your own
19:48:18 <frosch123> ogfx industries draws fences around industries, by detecting associated neighbouring tiles
19:50:30 <_dp_> frosch123, as childsprites you mean? won't that overdraw tile highlight?
19:51:45 <_dp_> nielsm, I feel bad doubling the amount of sprites in the game just for some silly eyecandy :p
19:52:26 <frosch123> _dp_: as secondary ground sprites
19:52:44 <frosch123> ground sprites can have child sprites themself, without adding a bounding box
19:54:30 <frosch123> worded differently. the highlight will be before the first building sprite, not after the first ground sprite
19:55:49 <glx> michi_cc: from command line regression works for me using project files (cmake, cmake --build ., ctest -C Debug .)
19:56:04 <_dp_> frosch123, oh, that's an interesting idea indeed
19:56:40 <_dp_> frosch123, how does it work though, will it render n sprites each time or compose them into same 5000 ?
19:57:03 <frosch123> it renders all sprites on their own
19:57:41 <_dp_> frosch123, so it kinda hurts rendering performance then?
19:58:36 <frosch123> you have only one action 123 chain for each tile
19:58:44 <frosch123> and regular tree tiles also draw one sprite for each tree
19:59:07 <frosch123> so, why should your objects be more expensive than other other objects or standard trees?
19:59:41 <_dp_> frosch123, yeah, sounds about right
19:59:48 <_dp_> guess I'll try that then
20:01:16 *** gelignite has joined #openttd
20:03:51 <_dp_> also probably can split tile in 4 and optimize amount of slopes it needs
20:17:11 *** HerzogDeXtEr1 has joined #openttd
20:36:52 *** Wormnest has joined #openttd
21:55:57 <glx> michi_cc: tried regression project inside VS and it also works, so I don't see what can be wrong (weird thing being Ninja doesn't show the issue for you)
22:05:23 <glx> michi_cc: but I can confirm the lang files issue
22:22:05 <michi_cc> glx: If I let cmake write REGRESSION_RESULT directly back into a file, I see that the last line ("dbg: [script] The script died unexpectedly.") is not terminated by an end-of-line character. result.txt on the other has an EOL at the end.
22:22:46 <michi_cc> It is entirely possible that the EOL of the regression output depends on CMake version, VS version, Windows version and moon phase.
22:23:35 <glx> but result EOL is sent by openttd
22:24:33 <michi_cc> It has one if I remove the ERROR_STRIP_TRAILING_WHITESPACE directive from execute_process.
22:28:30 <michi_cc> I've generated the project files with CMake 3.17.3. Not sure which cmake version is shipped with VS itself, which would be the version used in Open Folder/ninja.
continue to next day ⏵
