IRC logs for #openttd on OFTC at 2025-01-11
⏴ go to previous day
03:06:59 *** Wormnest has quit IRC (Quit: Leaving)
03:40:04 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:00:57 *** gnu_jj_ has joined #openttd
04:04:21 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:41:56 <DorpsGek> - Update: Translations from eints (by translators)
05:52:35 *** APTX has quit IRC (Ping timeout: 480 seconds)
07:40:02 *** kuka_lie has joined #openttd
08:06:43 <peter1138> Oh, I have some badge stuff in NML from September.
08:31:08 <peter1138> make: time: No such file or directory
08:31:08 <peter1138> make: *** [Makefile:108: generated/iron-horse.nfo] Error 127
08:37:56 <andythenorth> What did we break? 😮
08:39:15 <peter1138> `time` is normally just a built-in shell command on Linux systemn.
08:39:40 <peter1138> apt install time :S
08:40:00 <peter1138> Building iron-horse as GRFv9.
08:40:16 <peter1138> As it takes you a minute, it'll probably take me about an hour.
08:46:10 <peter1138> Parsing ... 118.0 s
08:46:10 <peter1138> Preprocessing ... 29.3 s
08:46:10 <peter1138> Generating actions ... 5.2 s
08:46:10 <peter1138> Assigning Action2 registers ... 224.2 s
08:46:15 <peter1138> And then it crashed.
08:46:57 <peter1138> nmlc info: Concurrent spritegroups: 203/32768 ("generated/iron-horse.nml", line 29085)
09:00:19 <LordAro> peter1138: it's builtin with bash, but not with dash ;)
09:04:03 <peter1138> Okay, I think I've found the errant `3`.
09:12:15 <peter1138> -rw-r--r-- 1 petern petern 42181114 Jan 11 09:09 iron-horse.grf
09:12:15 <peter1138> -rw-r--r-- 1 petern petern 89694598 Jan 11 09:08 iron-horse.nfo
09:12:15 <peter1138> -rw-r--r-- 1 petern petern 88818950 Jan 11 08:30 iron-horse.nml
09:13:31 <peter1138> Hmm, current released grf is 41641106
09:18:52 <andythenorth> my build of latest commit on `main` is 41,666,676 bytes
09:19:01 <andythenorth> just as another data point 😛
09:22:53 <andythenorth> peter1138: also why is it so slow 😛
09:24:54 <peter1138> [iron-horse.grf:11374] LoadNewGRFFile: Unexpected sprite, disabling
09:26:24 <peter1138> Hmm, maybe the goto label handling isn't right.
09:28:02 <andythenorth> if you build iron-ibex or iron-moose instead, they're a lot smaller, but will still exercise the majority of code paths
09:29:38 <andythenorth> ibex particularly I think
09:33:32 <peter1138> Yeah goto label handling was wrong.
09:58:52 <peter1138> Well, it was, but its not using labels here. Hm.
10:02:55 *** Smedles has joined #openttd
10:21:20 <peter1138> I *think* I worked it out.
10:33:16 *** kuka_lie has quit IRC (Ping timeout: 480 seconds)
10:36:32 *** kuka_lie has joined #openttd
10:37:31 *** akimoto has joined #openttd
10:41:50 <peter1138> Hah, action 6 offsets.
10:44:54 <peter1138> id, offset = actionD.write_action_value(id, action_list, act6, 5, 2)
10:45:00 <peter1138> I guess this is... somewhat fragile.
11:00:17 <xarick> How should I detect solo pockets of rivers
11:07:26 <peter1138> Ok, so reading past end-of-sprite... could be the sprite is wrong, or more likely the action 6 just before it is wrong. :S
11:08:07 *** Flygon has quit IRC (Read error: Connection reset by peer)
11:11:25 <peter1138> offset += parser.var_list_size + 1 # +1 for the byte num-ranges
11:40:50 <xarick> it's the last thing needing fix
11:50:02 <peter1138> Turn it into a feature.
11:53:41 <xarick> I'm more inclined into just cleaning them on a final check
11:54:30 <xarick> ideally these pockets shouldn't be created
11:55:34 <peter1138> Hmm, working out where these offsets are created is not fun.
12:28:52 <peter1138> Not got it yet. It's modifying byte 20 in both old and new code.
12:29:00 <peter1138> Should be byte 21...
12:29:45 <peter1138> Anyway, it turns out that during the initial grf loading, we skip Act8. So the grf_version isn't loaded.
12:30:26 <peter1138> So all the incompatible changes don't work.
12:31:41 <peter1138> I fixed it by not skipping the Act8. But this actually means that before the Act8, all handling is/should be treated as GRFv... something ancient.
12:32:34 <peter1138> Action 8 has to be one of the first pseudosprites in the .NFO file. It is however valid to have other actions before it, except those that define new sprites or vehicle (etc.) properties.
12:32:38 <peter1138> Actions 6, 7, 9, B, C, D (except resource management sprites), 10 and 14 are valid before an action 8.
12:33:14 <peter1138> Does NML ever output anything other than Action 14 before the Action 8...?
13:05:35 <_glx_> Some 6,7,B are possible (depends on nml file though)
13:06:43 <_glx_> Minimum openttd version checks
13:07:24 <LordAro> peter1138: it's good fun out there today
13:14:32 *** Tirili has quit IRC (Quit: Leaving)
13:20:58 <johnfranklin> Is it lunchtime?
13:36:21 <peter1138> LordAro, I didn't dare. Way too cold, I've not been acclimbatising over Autumn.
13:37:11 <peter1138> Of course my trike is suitable for it...
13:38:10 <peter1138> Ah yes, Action 6 still.
13:52:01 <LordAro> peter1138: difficult to fall over on a trike
13:52:02 <peter1138> _glx_, yeah, so if you writing a GRF you need to mix versions up until Act8.
13:52:33 <andythenorth> LordAro: challenge accepted?
13:52:36 <andythenorth> think you could roll it
13:53:55 <peter1138> 19565 * 7 06 7B 04 FF \wx0014 FF
13:54:15 <peter1138> 19561 * 6 06 7B 04 \wx0014 FF
13:54:27 <peter1138> That 0x0014... is not right.
13:54:32 <_glx_> hmm nmlc side will be a pain to handle
14:00:01 <_glx_> or nmlc could put a static forced version check at start, then all other actions should be fine using the new format
14:00:55 <peter1138> Only if they are after the Act8.
14:03:37 <peter1138> 19562 * 29 02 00 \wx7FE6 89
14:03:54 <peter1138> Also, because NML allocates set-ids from the top, it starts at 0x7FFF :D
14:04:46 <_glx_> but if we put a hardcoded act7 and actB in begin of grf
14:09:18 <peter1138> Hmm, were there any pre-GRFv2 files in existence...
14:16:35 <_glx_> doc says 0, 1, 4,... no v2 🙂
14:17:45 <_glx_> openttd doesn't support pre-v2 anyway
14:23:59 <peter1138> No, that was just idle curiosity :)
14:27:10 <_glx_> first ttdpatch release with official grf support was already using v2
14:29:45 <peter1138> Hmm, maybe it's because I did not delete the old version of NML from /usr/local/lib...
14:30:11 <peter1138> It's a bit frustrating that different versions are installed alongside each other, and it doesn't use the one I want :p
14:31:27 <peter1138> I'm just guessing now :(
14:31:40 <peter1138> Action 6 magical number processing.
14:33:43 <_glx_> yeah act6 is magic in nmlc 🙂
14:42:29 <peter1138> Difficult to see where things are coming from.
14:43:06 <_glx_> some times it's "direct" act6.modify_bytes() calls, sometimes is temporary storage in a returned variable to be passed to act6.modify_bytes
14:44:28 <_glx_> but searching for modify_bytes() uses should help to find where they come from
15:16:01 <_glx_> that could be a problem indeed
15:16:14 <_glx_> and no compiler complained about that ?
15:20:48 *** akimoto has quit IRC (Remote host closed the connection)
15:30:03 <_glx_> MSVC is usually a specialist for these warnings (but mostly for size_t to uint32_t in x86 builds)
15:30:41 <_glx_> or x64, but always one of them warns
15:33:06 *** Wolf01 is now known as Guest5623
15:34:49 *** Guest5623 has quit IRC (Ping timeout: 480 seconds)
15:39:59 *** Wormnest has joined #openttd
16:15:48 <peter1138> Okay, iron-ibex is working.
16:16:14 <peter1138> Let's give iron-horse 10 minutes to build.
16:18:26 <andythenorth> it's probably slow loading into your new game btw
16:18:31 <andythenorth> [unrelated channel chat]
16:19:15 <peter1138> My new game of building IH from source?
16:19:33 <andythenorth> apparently Horse makes new games slow to start
16:20:34 <peter1138> Eh, it's faster now.
16:24:28 <peter1138> nmlc info: Concurrent spritegroups: 203/32768 ("generated/iron-horse.nml", line 29085)
16:24:47 <peter1138> Is a 32-bit callback result desirable?
16:25:17 <peter1138> The only thing I can think of that would benefit is articulated engine ID parts.
16:25:33 <andythenorth> unrelated, but the 15 bit, unsigned results still bite me
16:25:54 <andythenorth> sometimes I want to return -1 to things 😛
16:26:08 <andythenorth> there's some var I can use also I think
16:27:00 <peter1138> Hmm, test FIRS next?
16:28:11 <peter1138> /home/petern/src/firs/.chameleon_cache/properties_industry_f921664ff3d46175ff254331e492d938.py:111: SyntaxWarning: "is" with a literal. Did you mean "=="?
16:28:39 <peter1138> FIRS is instant compared to IH.
16:30:56 <andythenorth> peter1138: try `v5-release-track` branch, warnings should be gone there
16:31:31 <peter1138> ModuleNotFoundError: No module named 'grf.incompatible_grfs'
16:32:10 <peter1138> Do I have to read things now? :p
16:32:50 <peter1138> Ok, FIRS is broken.
16:32:56 <peter1138> "Too many outputs (max 16)"
16:33:17 <peter1138> So I have probably forgot to change the size of something.
16:34:37 <andythenorth> sounds like a FIRS warning
16:37:21 <peter1138> Ah, the generic code means that the value for number of accepted/produced cargo types is 16 bits instead of 8 bits.
16:37:36 *** SigHunter has joined #openttd
16:38:31 <peter1138> In way, ditching extended byte is highlighting issues that would otherwise be hidden.
16:39:17 <peter1138> Although I fixed the property, but actually it's the callback that needs sorting. Hmm.
16:40:42 <peter1138> Who even asked for this?
16:48:20 <michi_cc> peter1138: Random though about callback results: Would there be any advantage in replacing the current `set-id` word with a `type` byte followed by a word with either the set-id or callback data? This way the meaning of the value would be explicitly coded instead of implicitly deducted.
16:49:08 <peter1138> So allowing set-id / callback to be fully 16 bit?
16:49:09 <michi_cc> And you could add a type for 32-bit return later 🙂
16:49:48 <michi_cc> And just make it explicit if the value is another action 2, a callback results, or a whatever we come up (maybe an explicit fail/end-of-chain marker).
16:51:07 <peter1138> But also I was wondering if somehow it should be even more breaking, and things become self-descriptive.
16:51:21 <michi_cc> How's return computed result currently implemented? Could be a separate type as well.
16:52:32 <peter1138> But adding "format" data to everything is probably not all that feasible and worthwhile... if you don't know what to do with it.
16:53:36 <peter1138> Okay, so result-type, instead of the bit 15 shenanigans.
17:17:53 <peter1138> I'll consider this stuff later.
17:18:08 <peter1138> Can I modify FIRS' build to output the NFO?
17:23:21 <andythenorth> change the makefile
17:23:43 <andythenorth> the Horse makefile shows the way
17:45:13 <peter1138> Well, beyond my Makefile ability right now.
17:48:08 <_glx_> just find the `nmlc` line to add `--nfo filename`
17:48:18 <andythenorth> hmm FIRS makefile is older
17:48:31 <andythenorth> inherited from coop
17:48:47 <andythenorth> need to modify `NML_FLAGS`
17:49:57 <_glx_> `--nfo generated/$(PROJECT_NAME).nfo` should do
17:50:14 <andythenorth> `NML_FLAGS =-c -l $(NML_LANG_DIR) --verbosity=4 --no-optimisation-warning --list-unused-strings --nfo=generated/firs.nfo`
17:50:38 <merni> The readme says nothing 😄
17:51:07 <_glx_> openttd is supposed to find a working font
17:51:25 <merni> never worked for me for Tamil
17:51:45 <merni> and that doesn't mean leaving in a wrong error message makes sense :p
17:51:48 <_glx_> ah maybe tamil font don't have full non tamil support
17:52:15 <merni> It does in that when I set a font manually that error no longer appears in the console
17:52:27 <merni> Yes some fonts don't have latin letters in them but many do
17:52:37 <merni> Including what I think is Windows's default Tamil font (Nirmala UI)
17:52:44 <_glx_> detection may not work properly
17:54:08 <_glx_> and fallback mechanism forcing all sizes to use the same font doesn't help
17:54:25 <peter1138> "leaving in a wrong error message"?
17:54:42 <merni> "Read the readme to see how to solve this"
17:54:50 <merni> The readme doesn't contain any information on how to solve it
17:55:18 <peter1138> Okay, that's not how I parse "leaving in a wrong error message"
17:55:31 <peter1138> That's "incorrect information in an error message"
17:55:56 <merni> perhaps "leaving in incorrect info in an error message" would be better
17:56:18 <peter1138> "leaving in" implies that somebody's been changing it but didn't bother to remove i.t
17:56:32 <merni> which is a better solution?
17:56:32 <merni> - add info to the readme on how to solve this
17:56:32 <merni> - add the same info directly to the error message
17:56:58 <merni> I prefer the second because players who install via, say, Steam won't necessarily know how to access the readme
17:58:02 <wensimehrp> the situation is even worse for players who use traditional chinese... 🙂
17:58:05 <merni> peter1138: I mean that when glx says "openttd is supposed to find a working font", that may be true, but it doesn't mean we can ignore the error message that happens when that fails
18:00:32 <merni> seems like the intent of that PR is essentially that the readme should be primarily an intro for people who come across the github page, rather than for players to read after installing?
18:00:58 <merni> In which case this would probably be better fixed by putting the guidance (such as it is) in the error itself
18:00:59 <_glx_> yeah we have wiki for users
18:01:27 <_glx_> and there's the `font` console command now too
18:02:26 <merni> yeah I guess that's what we should be telling people who see that error to use
18:02:34 <merni> In the absence of a GUI font picker :p
18:04:06 <merni> Because this error is a non-translatable string, there is also this fun situation when you select a font that doesn't have Latin letters
18:04:27 <merni> But I hope that doesn't ever happen by default (unless there are no Latin fonts on the system)
18:04:50 <merni> Hmm I guess even in that case you have the openttd font
18:05:00 <peter1138> So when loading NewGRFs, it seems it's possible that we don't have a _cur.grffile set, so we can't use the GRF version.
18:08:37 <merni> Hmm, the wiki search has ads :(
18:15:21 <merni> I get this when I submit an edit to a wiki page...
18:15:56 <merni> Even though I am evidently not "blocked" since I can access openttd.org fine up to that point
18:18:17 <merni> Yes as the page title says :)
18:18:22 <merni> What can I do to resolve this?
18:18:22 <merni> You can email the site owner to let them know you were blocked. Please include what you were doing when this page came up and the Cloudflare Ray ID found at the bottom of this page.
18:18:28 <merni> I should probably do that
18:18:51 <merni> But idk if email is a more effective way to reach whoever is responsible for this stuff than this channel 😆
18:19:24 <_glx_> maybe cloudflare doesn't like your IP
18:19:53 <merni> weird that it wouldn't take effect until I try to submit an edit, though
18:21:35 <peter1138> I guess I should've made a PR of that a while ago.
18:22:07 <peter1138> I can't be sure if it works correctly on all platforms.
18:22:57 <merni> could you do a PR build?
18:23:20 <merni> I want to test but I can't really be bothered to get windows compilation working lol
18:23:43 <peter1138> You used "we" a lot earlier, so I will say that "we" should be able to compile it, surely?
18:24:41 <_glx_> oh a PR build could help to get more testers
18:24:55 <peter1138> I can only trigger a preview build, which is not the same.
18:25:25 <_glx_> well macos fails so I won't trigger PR build
18:26:14 <peter1138> Probably a bad merge, I think it did the last time I tried :)
18:26:27 <_glx_> haha errors clearly show it was a blind edit
18:26:38 <merni> merni: this particular issue seems to be something different though, as there is actually a font that contains all the required characters, just openttd seems not to be able to find it
18:27:06 <merni> Will try to look at that
18:29:50 <peter1138> Well, maybe it never compiled.
18:29:55 <peter1138> And not just the merge.
18:30:19 <peter1138> I did have a self-PR to build it on the CI though.
18:30:34 <_glx_> but yeah the fail is more than just a typo
18:34:55 <peter1138> So I remember being annoyed at the way MacoS handles external glyphs, so I definitely looked at it.
18:35:04 <peter1138> But I guess I never actually got around to doing most of the changes.
18:35:57 <_glx_> it's a pain to write "untestable" code
18:36:39 <_glx_> even if it passed compile stage it could horribly fail
18:37:01 <peter1138> Also, Tamil works for me on Linux.
18:38:46 <merni> I am not surprised linux handles it better than windows
18:41:20 <_glx_> dbg: [fontcache:0] Font is missing glyphs to display char 0xBB5 in medium font size
18:41:54 <_glx_> nice it fails for a tamil letter
18:41:59 <peter1138> Also it is a draft for a good reason :)
18:42:43 <peter1138> I was testing the Windows build with mingw/wine.
18:42:57 <peter1138> So it does at least do something, but of course that isn't the same as real Windows.
18:43:41 <merni> Seems the only languages which fail for me are Hindi, Marathi and Tamil
18:43:47 <merni> Which are all Indic languages
18:44:04 <merni> So perhaps something in the way Windows handles Indic fonts is weird?
18:44:24 <_glx_> the way windows handles fonts is weird
18:45:00 <peter1138> For OpenTTD these fonts need to have the latin characters too.
18:45:12 <peter1138> Hindi is largely untranslated.
18:45:28 <merni> peter1138: Yes, but several tamil fonts do
18:45:44 <_glx_> ```[2025-01-11 19:45:23] dbg: [fontcache:0] Missing 46 glyphs in medium font size
18:45:44 <_glx_> [2025-01-11 19:45:23] dbg: [fontcache:0] Missing 36 glyphs in small font size
18:45:44 <_glx_> [2025-01-11 19:45:23] dbg: [fontcache:0] Missing 38 glyphs in large font size``` using the PR
18:45:54 <peter1138> Doesn't mean our font detection is finding the correct fonts.
18:46:05 <merni> That is what I am looking at right now
18:46:26 <merni> Though I probably won't be able to test properly without getting a windows compile working, bah
18:47:17 *** ChanServ sets mode: +v tokai
18:47:20 <peter1138> Fallback font detection is basically the same with this PR, so it's reasonable that it still fails to find a font.
18:49:06 <peter1138> And tbh, I'm not sure it even does anything right now for Windows.
18:49:50 <michi_cc> Well, we are only looking at the self-reported locales of the fonts for pre-selection. So if either the indic fonts or our translations don't have the right locale set, we just throw all fonts out.
18:50:50 <peter1138> With this PR, font_unix.cpp emplaces a new setting when it finds a font.
18:51:12 <merni> how do you check what locale a font has
18:51:16 <peter1138> font_win32.cpp does not. So yeah, that bit is definitely broken.
18:51:44 <peter1138> I think my Windows testing was probably "is it even able to load a font at all..."
18:53:12 <michi_cc> merni: How to do that easily is a good question 🙂
18:54:16 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:03:20 <peter1138> Multiple comments, I'm honoured :p
19:03:43 <peter1138> Anyway, some of that is from May. It's been sat around.
19:03:57 <peter1138> What other stupid things have I got?
19:24:36 <truebrain> merni: Always feel free to ping me with these kind of reports 🙂
19:24:56 <truebrain> A combination of country-of-origin and the posting of something that smells like a HTML injection, made Cloudflare go: let's not process this request.
19:26:25 <merni> maybe I should be sad that India evidently has a bad reputation with cloudflare's system
19:27:23 <truebrain> the good thing is, it means it is something in the content of the page you tried to update that is causing it 🙂
19:27:30 <truebrain> it is not just a flat-out: BAD
19:28:17 <truebrain> Yeah, it is the kinda page that I can imagine being detected as injection-ish 🙂
19:28:58 <merni> and now that I think of it I was using text like `[<font name>]` and `[<size>]` which "look like HTML" 😆
19:29:34 <truebrain> in general I have noticed that the WAF of Cloudflare doesn't really understand these kind of upload bodies
19:29:36 <merni> especially considering how easy it is to use a vpn
19:29:40 <truebrain> I have whitelisted many rules over the years 😄
19:30:13 <truebrain> I disabled whole of OWASP, for example
19:30:19 <truebrain> just because we do sanatize our user data correctly
19:30:22 <truebrain> but they don't know that 😄
19:30:33 <truebrain> right, try editing that page again if you like
19:30:36 <truebrain> that rule is now too whitelisted
19:32:23 <merni> I was able to make the edit and mess up the table markup
19:33:16 <truebrain> there is this Preview button you could use 😉
19:35:58 <merni> idk why the error was happening but it isn't anymore so I say problem solved
19:36:42 <truebrain> because I whitelisted the ruleset? Dunno, seems pretty obvisous why it was solved 😛
19:36:58 <merni> No, I mean the markup problem lol
19:37:14 <merni> Thanks for fixing the cloudflare issue, obviously that was you
19:37:18 <truebrain> that was also very obvious, you had a `|` you wanted to escape
19:37:35 <merni> Well I didn't realise I had to do that 😆
19:37:48 <truebrain> `|` is a table marker, as the first letter of that line suggests 😄
19:38:08 <merni> I just copied the previous row and edited what I wanted to say
19:38:18 <truebrain> learn wiki speak! 😛
19:38:35 <merni> I tend to use visual editor on wikipedia for anything somewhat complex
19:39:01 <truebrain> this new generation, not willing to learn anything 😛
19:47:10 <merni> Also, is the navigation bar at the top of wiki pages supposed to be vertically scrollable?
19:47:29 <merni> I forget what it's called. That thing with the "breadcrumb" in it
19:48:15 <merni> One side-quest at a time :P
19:48:38 <merni> Plus my CSS knowledge is worse than my wikitext
19:48:53 <andythenorth> Claude can CSS pretty reliably
19:48:59 <andythenorth> if you know what you want it to do
19:49:19 <andythenorth> humans shouldn't have been writing CSS ever really
19:49:35 <truebrain> what a weird thing to say
19:50:16 <truebrain> even Dreamweaver had something to say about that
19:53:22 <andythenorth> Dreamweaver was the second most disappointing application I learnt
19:53:27 <andythenorth> the first was Excel
19:53:39 <truebrain> ah, sorry, I forgot that Marvin entered the chat 🙂
19:54:27 <andythenorth> I drew this lovely website, thinking it was Photoshop
19:54:35 <andythenorth> squares and stuff
19:55:46 <andythenorth> I did a little course in Excel, I had never seen Windows before, the first thing I asked the teacher was 'where is the box where I type in what I want the computer to do and it does it'
19:56:02 <andythenorth> [mismatched expectations]
19:57:02 <merni> I thought Visual Basic and FrontPage were the coolest things
19:57:13 <andythenorth> I thought the web would be like InDesign or Quark Express 😄
19:57:45 <andythenorth> ok is it lunch yet?
19:58:48 <merni> No, it's past midnight snack time
20:13:03 <merni> > Linked page "en/Basesets" does not exist (wikilink "[[en/Basesets]]").
20:13:34 <merni> doesn't work even if I include the trailing slash in the link
20:17:30 *** gelignite has joined #openttd
20:18:45 <truebrain> read the instructions! `Main Page` is the index page 🙂
20:20:25 <peter1138> Hmm, okay, I haven't touched the production callback.
20:20:38 <peter1138> But it does have act6 stuff in it.
20:26:28 *** kuka_lie has quit IRC (Quit: Lost terminal)
20:27:24 <peter1138> No suitable places... that seems more like it.
20:27:46 <xarick> github copilot feature on visual studio is annoying
20:28:00 <peter1138> "(undefined string)", perhaps not.
20:28:18 <xarick> it suggests huge blocks of code at times
20:28:56 <xarick> 5 lines, 8 lines, pushing the other text down just to show the suggestion
20:31:03 <xarick> I wonder how much closed source code if being sniffed by these AI features...
20:31:54 <xarick> closed source espionage
20:33:20 <_glx_> ok let's try switching to GetLocaleInfoEx() (needs at least vista, should be fine)
20:46:20 <_glx_> probably should update that as we officially support 7+
20:48:26 <peter1138> Hmm, VS Code's diff is telling me every single line is different. Which is not true.
20:50:54 <peter1138> Ah, I think the files are too big :)
20:58:56 <truebrain> I had to double-check, but indeed all the hardcoded URLs are redirecting subdomains 🙂
21:04:36 <peter1138> So many things could be wrong.
21:04:55 <peter1138> Am I writing/reading the strings/callback/action6s correctly...
21:05:14 <peter1138> Where even in FIRS is it? :D
21:05:25 <peter1138> I wonder what NewGRF is a... bit smaller.
21:06:24 <andythenorth> you need industries?
21:09:10 <merni> Aha, I forgot about the Help section
21:10:11 <merni> That makes sense, though this error message would still need to reference either a file location or a URL if your font shows everything else as a bunch of boxes 😆
21:10:19 <merni> I will look at it tomorrow
21:11:12 <_glx_> of course when changing WINVER from XP to 7, some conflicts from API defines happen
21:34:23 <_glx_> just commented the line and
21:35:05 <_glx_> but of course now it tries all fonts
21:36:30 <_glx_> I'll see what I can do with fsUsb to prefilter
21:36:56 *** reldred has quit IRC (Quit: User went offline on Discord a while ago)
21:37:00 <_glx_> I guess something similar to fsCsb checks
21:37:01 <peter1138> If you fancy fixing up my PR so that the font(s) it finds are actually used...
21:37:36 <_glx_> I'm on master for now (to exclude other possible issues)
21:37:38 <peter1138> (Later, not yet :))
21:48:30 <xarick> ```inline void SetUnterraformableWaterState(Tile t, bool b)
21:48:30 <xarick> assert(IsUnterraformableWater(t) == !b);
21:48:30 <xarick> AssignBit(t.m3(), 1, b);
21:48:32 <xarick> question about the assert. How do I make it less weird
21:49:25 <_jgr_> `!= b` seems more obvious
21:50:09 <_jgr_> Though asserting that the new state is different to the old state is probably not very useful in general
21:51:53 <xarick> `assert(b == !IsUnterraformableWater(t));` hmm, slightly more elegant, but still weird overall
21:53:03 <xarick> I had a brain moment: `assert(b != IsUnterraformableWater(t));`
21:55:48 <xarick> is this a case of Cognitive decline? I feel embarrased
22:02:41 <peter1138> Hmm, well, the set-ids don't match.
22:06:12 <peter1138> I think I should expect them to be the same (except
22:12:44 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:36:17 <xarick> I got a false positive in TestRiverConnection
22:36:35 <xarick> but the fix makes it worse
22:37:12 <xarick> originally the terrain looks like that on the bottom left
22:37:39 <xarick> the mentioned tile is where the pathfinder is told to search
22:38:08 <peter1138> Are set-ids cached...
22:38:31 <xarick> but when the river is constructed, that tile becomes inclined land (top right image)
22:38:40 <xarick> TestRiverConnection fails
22:40:53 <xarick> marking the tile as unterraformable "fix" makes that river ending look much worse
22:43:41 <xarick> reverting the "fix", I don't like it
22:50:07 <xarick> must think of a better solution
22:55:38 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:56:56 <peter1138> Hah, I was correct, the Action 6 offset is wrong.
23:02:27 <_glx_> so easy to have wrong offset in act6
23:02:49 <_glx_> hard part is to find where the offset is determined
23:05:33 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:13:47 <peter1138> Yeah, in fact it looks correct after all. :(
23:14:00 <peter1138> I missed one \b -> \w in the output.
23:45:32 <peter1138> Ooh, figured it out.
continue to next day ⏵
