IRC logs for #openttd on OFTC at 2020-05-08
⏴ go to previous day
00:05:14 *** Smedles has joined #openttd
00:10:59 *** Smedles has joined #openttd
00:11:50 <DorpsGek_III> [OpenTTD/nml] FLHerne opened pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfcTE
00:14:39 <FLHerne> Oh, stupid circular imports
00:17:16 *** Smedles has joined #openttd
00:22:26 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfcTE
00:24:59 *** Smedles has joined #openttd
00:25:34 <DorpsGek_III> [OpenTTD/nml] Yexo approved pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfckY
00:26:08 <Yexo> Shouldn't be too hard to remove the copy/paste stuff by putting everything in a wrapper function
00:26:17 <Yexo> But that can also be done afterwards
00:26:22 <FLHerne> Yes, I actually wrote that
00:26:55 <FLHerne> Then I decided it was a bad idea to add stuff and change the implementation in the same change
00:27:11 <Yexo> I can only agree with that
00:27:35 <FLHerne> ...and then I thought "sod it, this entire file is a mess" and didn't bother pushing that last commit with the others :P
00:28:42 <DorpsGek_III> [OpenTTD/nml] Yexo commented on pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfckR
00:29:44 <FLHerne> And I should edit them onto the wiki
00:32:26 *** Smedles has joined #openttd
00:34:05 <DorpsGek_III> [OpenTTD/nml] glx22 commented on pull request #118: Fix: Allow calculation for palette in spritelayout https://git.io/JfckD
00:35:05 <DorpsGek_III> [OpenTTD/nml] FLHerne dismissed a review for pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfckY
00:35:05 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #119: Fix: Add missing `roadtype()` and `tramtype()` builtins. https://git.io/JfcTE
00:40:38 *** Smedles has joined #openttd
00:48:04 *** Smedles has joined #openttd
00:50:45 <DorpsGek_III> [OpenTTD/OpenTTD] James103 opened issue #8123: "Unable to find local depot" on dead-end tram tracks (regardless of location) https://git.io/JfcIn
00:55:45 *** Smedles has joined #openttd
01:01:28 *** Smedles has joined #openttd
01:03:29 *** WormnestAndroid has quit IRC
01:04:27 *** WormnestAndroid has joined #openttd
01:09:24 *** Smedles has joined #openttd
01:17:37 *** Smedles has joined #openttd
01:20:09 *** snail_UES_ has joined #openttd
01:24:20 *** Smedles has joined #openttd
02:52:48 *** Gustavo6046 has joined #openttd
02:54:07 *** Smedles_ has joined #openttd
03:19:18 *** Wormnest has joined #openttd
03:38:50 *** Smedles has joined #openttd
03:41:02 *** gelignite_ has joined #openttd
04:28:54 <DorpsGek_III> [OpenTTD/nml] jrook1445 opened issue #120: nml 0.5.0 - industries transported_last_month_pct variable not in range 0..100 https://git.io/Jfc3f
06:04:07 *** mindlesstux66 has joined #openttd
06:09:41 *** snail_UES_ has joined #openttd
06:51:59 *** sla_ro|master has joined #openttd
06:54:42 *** Foveafoxy has joined #openttd
06:55:18 *** Foveafoxy is now known as Guest24449
08:03:22 *** WormnestAndroid has quit IRC
08:03:35 *** WormnestAndroid has joined #openttd
08:17:54 *** WormnestAndroid has quit IRC
08:18:11 *** WormnestAndroid has joined #openttd
08:24:43 *** Smedles has joined #openttd
09:11:28 *** snail_UES_ has joined #openttd
09:20:57 *** Guest24449 has left #openttd
09:20:59 *** Guest24449 has joined #openttd
09:48:23 *** andythenorth has joined #openttd
10:12:11 *** snail_UES_ has joined #openttd
10:12:56 <andythenorth> couple of nice nml bug reports
10:38:31 *** JiMMY424 has joined #openttd
11:08:28 *** snail_UES_ has joined #openttd
11:19:32 *** matt21347 has joined #openttd
11:23:25 <FLHerne> andythenorth: Incidentally, your usage of "deprecated" continues to bug me
11:23:54 <FLHerne> "Deprecated" specifically *doesn't* mean "removed"
11:26:48 <nielsm> yes, it means "still exists but discouraged from use in new development"
11:27:14 <nielsm> (and encouraged to migrate existing systems off it)
11:28:13 <FLHerne> But all the nml commit messages say "deprecate *whatever industry stuff*" while deleting it all
11:30:56 <FLHerne> - percent_transported_passengers
11:31:10 <FLHerne> - last_month_transported
11:31:17 <FLHerne> - transported_last_month_pct
11:31:35 <FLHerne> (to be fair, that predates the new industry spec)
11:36:13 <rudolfs[m]> Speaking of "deprecated", yesterday I ran into nml bug, where it called a python function, that was previously deprecated, but now removed
11:40:50 <DorpsGek_III> [OpenTTD/nml] FLHerne opened pull request #122: Fix #120: Value range of transported_last_month_pct should be 0..100. https://git.io/Jfcux
11:41:16 <FLHerne> rudolfs[m]: That was fixed in 0.5.0
11:48:49 <rudolfs[m]> That's great! But I think it's still a good idea to take deprecations seriously, since packaged software tends to not be updated as frequently.
11:58:43 <FLHerne> rudolfs[m]: Your friendly Debian packager is blathijs, but AIUI he's waiting for 0.5.1 because of all the bugs :P
12:02:47 <andythenorth> FLHerne we could just change it to removed
12:02:54 <andythenorth> it's no skin off my nose
12:03:25 <andythenorth> nml wiki says 'removed' where appropriate
12:05:45 <andythenorth> FLHerne where are you finding 'deprecated'? I'll change them, unless its commit
12:06:24 <blathijs> FLHerne: Yup, the release tarball got a bit messed up with 0.5.0, so I'm indeed waiting for 0.5.1 which should make everything better :-)
12:06:43 <FLHerne> andythenorth: Commits mostly
12:07:03 <FLHerne> I think it was in an an earlier draft of the changelog, but gone in the final one
12:07:32 <FLHerne> (rewriting the git history of 0.5.0 just to fix that might be overkill :P)
12:07:43 <andythenorth> probably, but it's a note for future
12:07:48 <andythenorth> I agree on the terminology
12:07:59 <FLHerne> OTOH, you could squash all the random bugfixes and look more competent ;-)
12:08:20 <andythenorth> we seem to be holding open 0.5.1 for a while
12:08:25 <andythenorth> despite it being broken in prod
12:08:50 <andythenorth> if this was a work project I'd have released the early fixes
12:09:10 <andythenorth> on the basis that broken in prod is bad
12:09:25 <FLHerne> There are more fixes, though
12:09:25 <andythenorth> and now we have more emerging bugs, so we'll probably hold 0.5.1 for longer?
12:09:49 <andythenorth> I do find OpenTTD very confusing sometimes :)
12:10:34 <LordAro> depends how easy they are to fix
12:11:11 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #119: Add: Builtin functions `roadtype()` and `tramtype()`. https://git.io/Jfcgt
12:11:17 <andythenorth> I haven't looked yet, I need a day from my actual software day job :)
12:11:22 <FLHerne> The latest ones look trivial, I've just posted a PR for the first one
12:11:22 <andythenorth> I"m just hanging out for the chat
12:11:41 * andythenorth hasn't had a day off since Jan 2nd except public holidays
12:12:08 <FLHerne> I think #119 should go to stable? It's basically a fix for the NRT support
12:12:37 <FLHerne> (originally it was prefixed Fix:, but Yexo thought "Fix: Add..." was silly)
12:12:52 <andythenorth> does it have docs and some sort of test? :)
12:13:09 <FLHerne> I'll edit the wiki when it's merged
12:13:29 <andythenorth> I think that helps
12:13:35 <DorpsGek_III> [OpenTTD/nml] LordAro merged pull request #119: Add: Builtin functions `roadtype()` and `tramtype()`. https://git.io/JfcTE
12:13:38 <andythenorth> either test, or in examples
12:13:46 <andythenorth> wiki is very appreciated
12:14:13 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #122: Fix #120: Value range of transported_last_month_pct should be 0..100. https://git.io/JfcgG
12:14:27 <FLHerne> There are no roadtype tests in regression/ at all?
12:14:38 <andythenorth> I doubt it, my patch was clown shoes
12:14:43 <FLHerne> That makes it less easy to add to one :P
12:15:31 <DorpsGek_III> [OpenTTD/nml] LordAro dismissed a review for pull request #115: Fix #112: use setuptools_scm to determine version https://git.io/Jf3E2
12:16:40 <FLHerne> LordAro: 255 * 101 / 256 is 100
12:17:07 <LordAro> FLHerne: i figured it was something like that
12:17:17 <LordAro> i have a maths degree, don't y'know
12:17:23 <LordAro> long since forgotten everything
12:17:28 <FLHerne> The natural way would be 255/256 * 100, but you'd get 0 because integers :P
12:18:30 <DorpsGek_III> [OpenTTD/nml] FLHerne merged pull request #122: Fix #120: Value range of transported_last_month_pct should be 0..100. https://git.io/Jfcux
12:18:31 <DorpsGek_III> [OpenTTD/nml] FLHerne closed issue #120: nml 0.5.0 - industries transported_last_month_pct variable not in range 0..100 https://git.io/Jfc3f
12:19:05 <LordAro> i think we should just get on and release 0.5.1
12:19:38 <FLHerne> I think I'll just merge 119, writing regression tests for all of NRT is out of scope
12:21:18 <FLHerne> Hm, I could add it to the examples
12:21:33 <FLHerne> The examples aren't actually tested, are they?
12:21:50 <FLHerne> Perhaps they should be, at least that they build
12:22:14 <andythenorth> I think they should be
12:22:21 <andythenorth> I appreciate they're not conventional unit tests
12:22:33 <andythenorth> but they are a form of tested documentation
12:22:41 <andythenorth> and as we're sticking to the wiki (afaict)
12:22:48 <andythenorth> it's impossible to include docs with the PR
12:47:16 <DorpsGek_III> [OpenTTD/nml] FLHerne opened pull request #123: Docs: Add a use of `roadtype()` to the examples. https://git.io/Jfc2O
12:47:42 <FLHerne> (is that the right prefix? idk)
13:03:12 <andythenorth> catches me every time, I use Docs: elsewhere
13:05:14 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #123: Docs: Add a use of `roadtype()` to the examples. https://git.io/Jfc2O
13:06:16 <andythenorth> nobody makes black LEDs?
13:11:31 <rudolfs[m]> You mean just diodes?
13:12:02 <FLHerne> Diodes don't emit light
13:12:19 <Wolf01> I need a negative light one
13:13:21 <Wolf01> Nope, I need to make darker
13:15:40 <milek7> you just need portable black hole
13:16:01 <Wolf01> That sucks... literally :P
13:25:11 <andythenorth> isn't that self explanatory Samu? :)
14:17:52 *** Foveafoxy has joined #openttd
14:18:29 *** Foveafoxy is now known as Guest24477
14:44:34 <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #124: Fix #121, da57fb5: check StorageOp register range only once https://git.io/JfcK8
14:56:18 <DorpsGek_III> [OpenTTD/nml] glx22 updated pull request #118: Fix: Allow calculation for palette in spritelayout https://git.io/JfZi5
14:57:06 <DorpsGek_III> [OpenTTD/nml] glx22 commented on pull request #118: Fix: Allow calculation for palette in spritelayout https://git.io/JfcK7
15:30:07 *** andythenorth_ has joined #openttd
15:32:34 <FLHerne> I set up `make test` to test the examples
15:33:10 <FLHerne> But now the tests fail :P
15:33:56 <andythenorth_> that's...information :)
16:14:35 *** Wormnest has joined #openttd
16:26:27 <DorpsGek_III> [OpenTTD/nml] FLHerne opened pull request #125: Codechange: Build examples in regression testing https://git.io/JfcXX
16:28:53 *** Foveafoxy has joined #openttd
16:29:12 <FLHerne> At least some of the ANIM warnings seem to be bogus, or at least the pixel locations
16:29:31 *** Foveafoxy is now known as Guest24490
16:32:40 <glx> example_industry builds fine here
16:33:50 <FLHerne> glx: In that particular case, the --nml output is bogus, and the test-build of that fails
16:34:10 <FLHerne> If you remove that, something else crashes
16:38:51 <FLHerne> andythenorth_: You wrote that example - are test_produce_2 and _3 supposed to do anything?
16:38:57 <FLHerne> Or can I just delete them?
16:40:06 <glx> hmm unreferenced stuff should not be in --nml output
16:40:10 <FLHerne> (and what are the STORE_TEMPs for? nothing reads them)
16:40:28 <nielsm> I probably wrote that example
16:40:47 <nielsm> and I don't remember why it'd be incomplete like that
16:41:01 <FLHerne> nielsm: Ok, andy's down as the commit author
16:43:14 <nielsm> looking at it, I think that produce might just be wrong
16:43:34 <nielsm> it will loop producing infinite STEL
16:44:21 <nielsm> (is there a requirement the examples are meaningful when loaded in the game?)
16:46:13 <FLHerne> I think so, otherwise they're not sensible examples?
16:46:58 <FLHerne> An example of how to produce infinite cargo is useful to no-one and misleads everyone else?
16:47:17 <FLHerne> I agree, that seems to be what it does
16:47:35 <glx> Produce2 __str__ is wrong anyway
16:48:24 <glx> or it's some __str__ in the elements
16:52:24 <glx> ha no, it's Produce2 __str__, it doesn't handle lists properly
16:55:55 <FLHerne> Why does Produce not-2 still exist?
16:56:09 <FLHerne> I don't think it's used anywhere
16:56:28 <FLHerne> nielsm: ^ is that left over from when you supported the old syntax too?
16:58:27 <glx> looks like 0.4 syntax yes
17:01:50 <andythenorth_> FLHerne where am I looking for test_produce_2 and test_produce_3?
17:02:05 <andythenorth_> I don't see them in examples/industry/example_industry.nml
17:03:32 <andythenorth_> I've considered writing a full set of examples
17:03:36 <andythenorth_> as a reference implementation
17:03:44 <andythenorth_> but I spend so much time on newgrf tools now
17:03:50 <andythenorth_> I can't write any of my own newgrfs
17:03:59 <andythenorth_> there's a probably a catchy name for that :P
17:10:21 <FLHerne> glx: Are you still looking at the Produce2 printing, or should I?
17:11:37 * FLHerne tries to fix the tracktypetable.__str__ crash
17:14:16 <glx> and I think Produce2 should be renamed Produce, and the old one can go
17:16:18 <glx> I could add the cleanup in the PR
17:18:49 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick opened pull request #8124: Fix f538179, #8119: Update docking tile area when placing a diagonal rail next to a dock end https://git.io/Jfcyk
17:20:40 <Samu> why's the commit checker complaining ? :(
17:21:11 <glx> commit checker should tell you where the error is
17:21:39 *** supermop_Home has joined #openttd
17:22:09 <glx> oh start with issue number
17:22:33 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick updated pull request #8124: Fix f538179, #8119: Update docking tile area when placing a diagonal rail next to a dock end https://git.io/Jfcyk
17:22:54 <andythenorth_> that's just a tiny part
17:24:18 <FLHerne> (if you've not seen it)
17:24:32 <andythenorth_> it also disturbs me they had FIRS 1.4.5 on their server for so long
17:24:34 <FLHerne> Those online docs are great, and also FIRS is great
17:24:39 <andythenorth_> I nearly joined reddit to ask them to stop
17:25:38 <andythenorth_> should I do natural gas as a cargo, or have a gas cracker that acts like a primary?
17:25:57 <andythenorth_> i.e. it produces the downstream products of natural gas with no supply
17:37:36 *** snail_UES_ is now known as Guest24498
17:37:36 *** snail_UES_ has joined #openttd
17:50:59 <FLHerne> I kind of liked FIRS 1
17:51:06 <FLHerne> The clustered industries were nice
17:56:18 <andythenorth_> I remember those
17:56:39 <andythenorth_> if cargo was evenly distributed to industries from stations, I might have kept them
17:56:49 <andythenorth_> but the mechanic was broken, not in a good way
18:08:30 *** gelignite has joined #openttd
18:12:25 <glx> FLHerne: tracktypetable.debug_print() is wrong too, but at least it doesn't crash
18:15:50 <supermop_Home> andythenorth_ but little RVs to deliver supplies were cute
18:22:31 *** Wormnest has joined #openttd
18:24:20 <FLHerne> glx: Meh, I think debug_print is pointless and redundant anyway, but I'll fix it :P
18:26:37 <andythenorth_> too many times they can't fit supermop_Home
18:26:41 <andythenorth_> overlapping delivery stations
18:26:45 <andythenorth_> I could probably have tuned that
18:26:51 <andythenorth_> but it was frustrating
18:42:59 * FLHerne checks the various ANIM warnings
18:55:39 <FLHerne> (the ones in the object example were actual stray pixels, I think)
19:00:09 *** Progman has joined #openttd
19:00:50 <FLHerne> andythenorth_: Can you review it? Then I can merge it and rebase the testing one
19:07:21 <andythenorth_> I am a bit afk sorry :)
19:13:36 <DorpsGek_III> [OpenTTD/nml] FLHerne approved pull request #124: Fix #121, da57fb5: check StorageOp register range only once https://git.io/Jfc52
19:19:35 <DorpsGek_III> [OpenTTD/nml] FLHerne merged pull request #124: Fix #121, da57fb5: check StorageOp register range only once https://git.io/JfcK8
19:27:38 <FLHerne> glx: Typo in the title
19:30:18 *** frosch123 has joined #openttd
19:33:54 <FLHerne> nielsm: I tried the example industry in-game, on first delivery it throws up "'Factory' caused an endless loop in the production callback" as you suspected
19:34:18 <frosch123> is it possible to add a custom text to the "merge" button on github?
19:34:20 <FLHerne> (I was curious if it was going to crash the game)
19:34:33 <frosch123> people keep on forgetting to update the docs when they add stuff to nml
19:34:56 * FLHerne realizes he forgot to update the docs :P
19:39:36 <andythenorth_> frosch123 that's the main reason I wanted to ditch the wiki
19:39:47 <andythenorth_> I am convinced it's not a good reason
19:39:55 <andythenorth_> does GitHub have custom workflows?
19:40:11 <frosch123> andythenorth_: people also forget to add stuff to the ai/gs changelog in ottd
19:40:40 <andythenorth_> at work we have deployment steps for tickets
19:40:45 <glx> and sometime forget to run the generate scripts
19:40:47 <frosch123> so unless you manage to add an automatic commit check... just putting it into the same repo does not solve it
19:41:02 <andythenorth_> a ticket can be merged but not deployed, so we have a stage for that
19:41:11 <andythenorth_> I am -1 to lots of workflow, but it has uses
19:56:33 *** ChanServ sets mode: +v tokai
19:56:51 <andythenorth_> hmm not sure 'workflow' means the same thing in github
19:57:19 <andythenorth_> seems to be GH Actions
19:57:45 <frosch123> we can make a PR template with a list of things to remind about
19:58:08 <FLHerne> Can someone just click 'accept' on #128? All it does is add ANIM flags and change two pixels
19:58:10 <andythenorth_> we could make a merge automatically make a new nagging issue to update docs :P
19:58:15 <andythenorth_> FLHerne I will yes 1 min
19:58:33 <FLHerne> Trying to test it, I'm puzzled that the object can't actually built in-game...
19:58:54 <FLHerne> It says "Can't build object... Infrastructure"
19:59:05 <FLHerne> (same is true before)
19:59:07 <andythenorth_> GH has image diff :)
19:59:13 <andythenorth_> too small to see anything but still nice
20:01:15 *** D-HUND is now known as debdog
20:01:45 <FLHerne> frosch123: Any idea what "Can't build object... <classname>" means? I wondered if INFR was a special-case, but it's the same if I change it
20:02:41 <frosch123> nml examples often use dummy texts, so <classname> probably means nothing
20:02:49 <frosch123> is there a location check in the nml?
20:03:57 <FLHerne> frosch123: In-game, I mean
20:04:21 <FLHerne> The newgrf compiles fine, but the object it defines can't be built
20:04:43 <frosch123> tile_check: return 0; <- that disallows unconditionally :)
20:05:36 <frosch123> and uses the first string as error message, whichever that is
20:07:42 <FLHerne> This seems like a bad example :P
20:07:57 <FLHerne> The first string is the class name, which is why it's that
20:11:26 <frosch123> please do, i never looked at the example :)
20:11:54 <andythenorth_> another day I am happy to help with examples
20:12:03 <andythenorth_> today I really need a day off from 'making mistakes with code'
20:12:10 <andythenorth_> which I have been doing for like 4 months no break
20:12:40 <andythenorth_> my hypothesis: if we have good examples, we can then ask for PRs to include examples
20:12:48 <andythenorth_> which will stop people asking me to write them
20:13:25 <andythenorth_> actually I'm exaggerating the 2nd point, but still I think PRs should come with an example so reviewer can validate
20:14:35 <frosch123> andythenorth_: the other day someone uploaded a industry newgrf to bananas, and credited manual industries as source, despite adding dozen of cargos. maybe some day people will fork the nml example for their stuff :p
20:15:56 <andythenorth_> reddit wants a daylength newgrf
20:19:32 <frosch123> reddit got excited that there is more than one firs
20:19:54 <frosch123> it was not quite clear, whether they were refering to firs 1/2/3, or to economies
20:20:45 <frosch123> hmm, we are at firs4 now?
20:27:45 <_dp_> oh waw, openttd now opens saves even when missing newgrfs?
20:27:55 <_dp_> how's that not breaking everything?
20:28:23 <andythenorth_> it falls back to stuff
20:28:33 <andythenorth_> it's kind of essential for newgrf dev btw
20:28:58 <frosch123> _dp_: you probably enabled newgrf dev settings
20:29:11 <frosch123> or one of the other hidden magic ones
20:29:41 <FLHerne> (did DorpsGek go away?)
20:30:10 <DorpsGek> Samu: I have not seen DorpsGek.
20:30:39 <FLHerne> DorpsGek: Get yourself a mirror, you silly bot
20:30:39 <DorpsGek> FLHerne: Error: 'yourself' is not a valid topic number.
20:32:56 <frosch123> FLHerne: enable whitespace in your editor
20:33:38 <DorpsGek_III> [OpenTTD/nml] frosch123 commented on pull request #130: Fix: Enable the example object to be placed in-game https://git.io/JfcNe
20:34:09 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #130: Fix: Enable the example object to be placed in-game https://git.io/JfcbS
20:34:13 <frosch123> well, looks like the whole file is inconsistent :)
20:34:42 <DorpsGek_III> [OpenTTD/nml] frosch123 approved pull request #130: Fix: Enable the example object to be placed in-game https://git.io/JfcNU
20:37:13 <DorpsGek_III> [OpenTTD/nml] FLHerne merged pull request #130: Fix: Enable the example object to be placed in-game https://git.io/JfcbS
20:46:00 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #125: Codechange: Build examples in regression testing https://git.io/JfcXX
20:46:14 <FLHerne> ^ the tests should actually pass now, thanks glx!
20:47:16 <FLHerne> nmlc doesn't have a "fail on warning" flag, does it?
20:47:43 <FLHerne> They probably *shouldn't* actually pass, because the industry example is still broken
20:48:36 <frosch123> it has the opposite :) -q suppresses warnings
20:49:29 *** snail_UES_ has joined #openttd
20:49:29 <FLHerne> Why did anyone think that was a good idea?
20:49:43 <FLHerne> Huh, the tests failed, why?
20:50:36 <FLHerne> Oops, didn't rebase properly after the last commit
20:53:46 <Samu> 1.8.0 was the only 1.8 version right?
20:53:53 <DorpsGek_III> [OpenTTD/nml] FLHerne updated pull request #125: Codechange: Build examples in regression testing https://git.io/JfcXX
20:57:02 <FLHerne> Yay, the example tests pass!
20:57:20 *** snail_UES_ is now known as Guest24516
20:57:21 *** snail_UES_ has joined #openttd
21:03:08 <TrueBrain> FLHerne: I guess that makes it beer-time!
21:10:31 <FLHerne> We should have a town_names example
21:11:08 <frosch123> use an existing grf then
21:11:16 <rudolfs[m]> I used Lithuanian repo as a template
21:11:18 <frosch123> with an grfid that is already used on bananas
21:11:35 <rudolfs[m]> Speaking of which, how do I get my town names into bananas?
21:11:43 <frosch123> townname people tend to copy stuff without understanding it, and then do not set their own grfid
21:12:00 <frosch123> rudolfs[m]: you go to bananas.openttd.org
21:12:07 <frosch123> everything else is written there
21:13:15 <FLHerne> Unrelated, it seems daft that everyone sets `if (version_openttd(1,2,0,22723) > openttd_version) {`, ... or so
21:13:26 <FLHerne> It's usually wrong or outdated anyway
21:14:03 <FLHerne> For all but very complex grfs, nmlc should know better
21:14:06 <frosch123> i recommend against any version checks, unless nml learns to add them itself
21:14:15 <FLHerne> (it doesn't, of course)
21:14:49 <frosch123> i also think checking revisions is too detailed. people should just check releases
21:15:55 <glx> checking revision was useful for test version during feature implementation in openttd
21:16:30 <glx> but since the move to github...
21:18:47 <frosch123> when you write the grf, you do not know when all bugs in ottd are fixed :)
21:18:54 <frosch123> so a revision is pretty useless
21:19:08 <frosch123> nightly players use nightly
21:23:44 *** Wormnest has joined #openttd
21:51:51 <Samu> how would you name the 3 variables pertaining to a dock?
21:53:47 <Samu> dock_end, dock_pier, dock_water, so many ideas, all of them not clear enough of their own
22:01:35 *** Laedek_ has joined #openttd
22:01:35 *** osvaldo[m] has joined #openttd
22:01:35 *** ookfof[m] has joined #openttd
22:01:35 *** nolep[m] has joined #openttd
22:01:35 *** dwfreed has joined #openttd
22:01:35 *** super_spooky has joined #openttd
22:06:39 <Samu> just got some ideas based from the sprites: docking_tile, dock_slope_tile, dock_flat_tile
22:12:05 *** dwfreed has joined #openttd
22:54:15 *** dwfreed has joined #openttd
23:06:13 *** Wormnest has joined #openttd
23:07:28 *** Progman_ has joined #openttd
23:17:41 <glx> FLHerne: for windows it should be easy to get zlib headers with vcpkg
23:21:45 <FLHerne> glx: I guess so, it's a question of how much we care
23:22:00 <FLHerne> We already test cross-platform with CPython
23:24:16 <FLHerne> So is it really worth downloading and compiling a bunch of stuff several times for every PR, for the sake of the 0.0whatever% of bugs that happen to affect only PyPy and only on certain platforms?
23:24:55 <FLHerne> (and actually get detected by our test suite at all...)
23:25:38 <FLHerne> I'd expect platform-related bugs to be about paths, weird encodings, whatever
23:25:44 <FLHerne> And we basically don't test those
23:26:10 <glx> oh for sure there are issues with encoding :)
23:27:10 <glx> on my machine with "pip install -e ." I can run nmlc, with "pip install ." I can't :)
23:27:50 <glx> but it's probably an issue with setuptools
23:29:21 <glx> oh and of course I can't run nmlc with "pip install nml"
23:32:28 <glx> for some reason "-e" uses "setup.py develop" I think, which uses easy install and generate a script with "#!<path>/python.exe -x" while standard install generates "#!<path>/python.exe" and fails to run because I have a ï in <path>
continue to next day ⏵
