IRC logs for #openttd on OFTC at 2022-06-15
⏴ go to previous day
00:00:06 *** wallabra has quit IRC (Ping timeout: 480 seconds)
00:45:55 *** wallabra has joined #openttd
01:29:10 *** Smedles has joined #openttd
02:22:09 *** Wormnest has quit IRC (Quit: Leaving)
02:43:57 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:39:47 *** andythenorth has joined #openttd
06:19:21 *** wallabra has quit IRC (Ping timeout: 480 seconds)
06:22:50 *** sla_ro|master has joined #openttd
06:29:07 <andythenorth> hmm I have done an evil thing
06:31:40 *** HerzogDeXtEr has joined #openttd
06:34:32 <andythenorth> I have some lists of prerequisites where every prerequisite in the list is built by a single python script, which is intended to be run once
06:34:50 <andythenorth> but make will run the script for each prerequisite
06:35:05 <andythenorth> in a non-parallel compile that's inefficient
06:35:29 <andythenorth> in a parallel compile it's causing errors as the two running scripts scribble over each other (e.g. adding and deleting dirs etc)
06:36:24 <andythenorth> I am sure there is a better way than having a no-op target for the individual nml files, but I don't know what it is
06:44:58 *** andythenorth has quit IRC (Quit: andythenorth)
07:17:10 *** andythenorth has joined #openttd
07:26:21 *** Smedles has joined #openttd
08:29:07 <andythenorth> I need to rewrite the python files to handle individual prerequisites anyway
08:29:12 <andythenorth> so this is a temporary shim
08:34:40 <andythenorth> eh, sys.argv relies on positional args
08:34:44 <andythenorth> this is limiting :P
08:34:56 <andythenorth> maybe I can use argparse
08:35:12 <LordAro> andythenorth: the "proper" way is to rewrite the python file as you say, but in the meantime you should be able to give the $(NML_FILES) rule a dependency of nml_cabbage, and make the nml_cabbage rule create the file (touch $@)
08:35:25 <LordAro> make is clever enough to only run the rule once then
08:35:41 <andythenorth> I think that's what the graphics target does
08:35:47 <andythenorth> frosch added a similar approach
08:57:48 <LordAro> bit warm for the next few days, for sure
08:57:55 <LordAro> unfortunately saturday is looking a bit damp
09:00:07 <peter1138> Today is the mid-week rides, at least for the retired/non-working members. Bah!
09:02:17 <andythenorth> non-working is the new in thing
09:02:37 <andythenorth> not sure how that squares with stagflation and cost of living, but eh
09:13:03 <andythenorth> so maybe I sort of understand makefiles
09:13:12 <andythenorth> after 10 years of using them :P
09:13:20 <andythenorth> this might be a false dawn but eh
10:35:40 <Samu> so i uninstalled visual studio community 2019 and now visual studio community 2022 can't build openttd...
10:59:21 <peter1138> Install Debian instead.
11:00:30 <LordAro> is that really the best idea here? :p
11:03:10 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
11:07:28 *** Eddi|zuHause has joined #openttd
12:08:23 <glx> do you get an error message Samu ?
12:10:18 <glx> but I think the first step would be to clear cmake cache (if you did not when switching to VS2022)
12:13:46 <Samu> i switched from x64 release to x64 debug
12:13:59 <Samu> then back to x64 release and it worked
12:14:27 <Samu> main menu had a background with water though
13:27:40 *** frosch123 has joined #openttd
13:33:14 <frosch123> hmmm, how do I edit pngs now... install some older gimp version? try to patch gimp myself? try some other software?
13:42:34 <andythenorth> frosch123 paint.net?
13:43:16 <andythenorth> oh that's not actually a thing, just a meme
13:43:48 <frosch123> i use that at work on windows
13:44:16 <frosch123> anyway, i think i found the correct gimp source
13:44:22 <frosch123> just need to dust of my gitlab account
13:45:02 <frosch123> gnome self-hosts a gitlab
13:45:08 <frosch123> so ofc i have no account there
13:45:48 <frosch123> apparently i can sign-in with gh and gl though
13:46:09 <frosch123> why do they host their own gitlab though?
13:47:39 <WormnestAndroid> frosch123: yes, that's me
13:49:53 <WormnestAndroid> Own gitlab I think is to have more control with deploying to websites,among other things.
13:53:19 <dwfreed> a lot of larger open source orgs have a social contract they need to adhere to
14:16:52 <frosch123> well... compiling gimp seems to be another story :)
14:17:11 <frosch123> it may be easier to use a hexeditor on the regular binary
14:24:17 <WormnestAndroid> As long as you use meson (master branch only) building shouldn't be too bad.
14:25:11 <frosch123> it wants a installation of babl and gegl, which i won't do without a container
14:25:28 <frosch123> and the Dockerfile claims to be abandoned
14:25:48 <frosch123> *dev version of babl/gegl
14:33:35 <andythenorth> frosch123 I edit PNGs with PIL :P
14:40:18 <glx> doesn't mean people will stop to use it :)
14:42:08 <frosch123> now building a gimp tag, which matches the system's gegl/babl
14:45:32 *** Wormnest has joined #openttd
14:49:29 <Wormnest> yeah, GIMP usually needs fairly up-to-date babl and gegl since they are closely tied together
14:52:46 <frosch123> great, now i cannot reproduce my issue :p
15:00:05 <frosch123> probably user error, i guess i did something that added an alpha channel
16:37:35 *** Smedles has joined #openttd
17:04:38 *** andythenorth has quit IRC (Quit: andythenorth)
17:22:12 *** Alkel_U3 has quit IRC (Quit: maintenance)
17:26:47 *** Alkel_U3 has joined #openttd
17:33:42 *** gelignite has joined #openttd
17:44:47 *** gelignite has quit IRC (Read error: Connection reset by peer)
17:45:12 *** gelignite has joined #openttd
17:46:25 *** andythenorth has joined #openttd
17:48:57 <andythenorth> time for adventures in argparse
17:55:32 <Samu> is it aproximate or approximate?
18:01:12 <andythenorth> I have args to makefile, which are consumed by make or passed on to other build scripts
18:01:25 <andythenorth> if I do normal -f or --foo= names
18:01:35 <andythenorth> that might collide with make built-in args
18:03:24 <frosch123> isn't there -- for that?
18:03:48 <andythenorth> that's maybe what I needed to know
18:03:50 <frosch123> also, makefiles with arguments, really? :p
18:03:57 <andythenorth> google is not finding me much
18:04:55 <andythenorth> they mostly relate to 'go faster' things
18:05:26 <andythenorth> one of them tells how many python pool workers to use
18:05:31 <andythenorth> could just abuse -j for that
18:06:03 <andythenorth> the only other one allows building one grf instead of all 3
18:06:14 <andythenorth> I would do a target, but I can't figure out how to combine targets
18:06:20 <andythenorth> e.g. `make install moose`
18:06:34 <andythenorth> literally what targets are for though, afaict
18:06:53 <andythenorth> maybe I can evaluate the passed targets
18:08:40 *** ChanServ sets mode: +o orudge
18:14:00 <andythenorth> googling for conditionals based on target name(s) gets nothing useful
18:14:14 <andythenorth> apart from a faint suggestion that it's a very weird idea, and use args instead
18:17:11 <frosch123> just use separate targets
18:17:23 <frosch123> install: install_pony install_ibex install_andy
18:17:52 <frosch123> build: build_pony build_ibex build_andy
18:17:59 <frosch123> build_andy: andy.grf
18:18:26 <frosch123> andy.nml: horse.pynml
18:20:32 <andythenorth> is there some way I can parse the target name?
18:20:46 <andythenorth> I know I can set the var in each target, but eh
18:21:19 <andythenorth> install_pony, html_docs_pony, release_pony, no_cargo_sprites_pony, no_docs_pony
18:22:00 <frosch123> why set the roster at all?
18:22:17 <frosch123> just generate all three nml each time
18:22:27 <frosch123> it's the nml step you want to save, right?
18:22:50 <andythenorth> I want to generate only one grf when developing
18:23:00 <andythenorth> everything is so slow, even with -j
18:23:16 <andythenorth> release is unknown so far
18:23:35 <frosch123> you also want to skip the pynml part?
18:23:48 <andythenorth> skip anything not related to this roster
18:23:58 <andythenorth> also I often use `make html_docs` for actual dev, because it's so much faster to develop via docs
18:24:35 <frosch123> maybe just use an environment variable then
18:24:43 <frosch123> export HORSE_ROSTER=moose
18:24:54 <frosch123> in Makefile: HORSE_ROSTER?=all
18:24:57 <andythenorth> I think I can drop some of the other args
18:25:06 <frosch123> so, it defaults to all, but you can set a single one in advance
18:25:07 <andythenorth> oh that's a nice shortcut
18:25:32 <andythenorth> I can probably abuse -j for the python pool workers
18:25:40 <andythenorth> and the other speed ups aren't really much faster maybe
18:27:26 <andythenorth> ok one of the speed ups is broken :P
18:27:30 <andythenorth> so that's useless :)
18:30:34 <TrueBrain> are OpenGFX+ repos migrated? Or lost in time now?
18:33:05 <frosch123> why one do you want?
18:33:14 <TrueBrain> issue of a bit earlier
18:33:53 <frosch123> ah, well, i am too lazy to migrate them
18:34:38 <TrueBrain> so the answer to the dude is: no, there is no location (anymore) :P
18:34:54 <TrueBrain> would you mind answering? :D
18:35:32 <frosch123> i assumed you wanted them as truegrf templates or so :)
18:37:19 <andythenorth> wonder how I should abuse -j
18:37:29 <andythenorth> maybe just pass it straight through and cross my fingers
18:49:59 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
19:17:48 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
19:17:48 *** Wormnest has quit IRC (Read error: Connection reset by peer)
19:18:02 *** WormnestAndroid has joined #openttd
19:18:04 *** Wormnest has joined #openttd
19:23:29 <glx> boto3 and botocore are noisy (well dependabot is)
19:28:06 <TrueBrain> there is a reason I didn't want it to announce here :P
19:30:11 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:30:11 <glx> at least it doesn't do all repo at once
19:30:51 *** ChanServ sets mode: +v tokai
19:37:38 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:45:32 <andythenorth> oh make has functions :P
19:45:36 <andythenorth> any reason not to use?
19:46:03 <andythenorth> need to do some repetitive processing of python args
19:47:06 *** WormnestAndroid has joined #openttd
19:48:00 *** andythenorth has quit IRC (Quit: andythenorth)
20:15:32 *** andythenorth has joined #openttd
20:30:12 <frosch123> andythenorth: do you know the rule "use python instead of shell scripts"?
20:30:28 <frosch123> it may have a sister "use python instead of make"?
20:32:12 <Samu> why can't i set my AI to require API 1.12?
20:33:54 <Samu> i get unknown error when compiling script
20:33:56 <andythenorth> my python graphics processor has a small adventure into 'only generate if newer'
20:34:00 <andythenorth> but I didn't enjoy it :P
20:34:31 <andythenorth> I am quite enjoying finally learning some make
20:34:39 <andythenorth> also the magic arguments are amusing
20:35:13 <glx> Samu: didn't check but 1.12 might be 12
20:38:43 <Samu> I am about to upload the new version of LuDiAI AfterFix
20:43:21 <Samu> still have to write the changelog, it's a tedious task
20:43:39 <Samu> i made too many changes at once
20:44:04 <Samu> but didn't write them down
20:48:13 <andythenorth> oh can't read make -j flag value :P
20:51:13 <andythenorth> echo --jobserver-fds=3,4 -j)
20:51:33 <andythenorth> not sure how I split -j out of that, or how to interpret the values
20:52:03 <andythenorth> I don't really understand shell args
20:52:10 <andythenorth> I assumed it would be available to shell somehow
20:54:47 <glx> an awk script could work too ;)
20:56:07 <glx> shell has access to command line arg, but make is not shell, it already parsed args from shell
20:58:18 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:04:22 <andythenorth> should I wrap make in a shell script? :P
21:04:29 <andythenorth> maybe also some perl?
21:10:03 *** wallabra has joined #openttd
21:11:07 <frosch123> show it to your CTO
21:11:35 <andythenorth> I like the Crackers
21:11:48 <frosch123> yeah, but i already had some other snack instead
21:11:54 <andythenorth> that is a superb diagram
21:12:07 <frosch123> i am surprised myself :)
21:12:35 <andythenorth> maybe I could run make via Apache and cgi-bin?
21:31:45 *** sla_ro|master has quit IRC ()
21:32:13 <Samu> erm, the auto-fill option in bananas upload form can be a bit annoying, I can't edit some text descriptions or tags
21:32:29 <Samu> I added ship support, and wanted to change that part a bit
21:47:32 *** D-HUND is now known as debdog
21:48:29 *** andythenorth has quit IRC (Quit: andythenorth)
21:52:51 <Samu> does it run? someone test ludiai afterfix v19
21:54:31 *** Samu has quit IRC (Quit: Leaving)
21:59:24 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:57:15 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:11:06 *** gelignite has quit IRC (Quit: Stay safe!)
continue to next day ⏵
