IRC logs for #openttd on OFTC at 2025-07-12
⏴ go to previous day
01:01:21 *** kuka_lie has quit IRC (Quit: Lost terminal)
02:17:30 *** Wormnest has quit IRC (Quit: Leaving)
02:54:25 *** tokai|noir has joined #openttd
02:54:25 *** ChanServ sets mode: +v tokai|noir
03:01:24 *** tokai has quit IRC (Ping timeout: 480 seconds)
03:03:07 *** gnu_jj_ has joined #openttd
03:06:21 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
06:24:43 <LordAro> i guess my check annotations change a few months back was not perfect
06:26:08 <LordAro> ah, this one is an info level annotation, not a warning
06:46:12 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
07:07:03 *** dh1 has quit IRC (Ping timeout: 480 seconds)
08:17:00 <peter1138[d]> 2 years later though
08:20:12 <andythenorth> OS/2 on Apple Silicon?
08:24:41 *** dh1 has quit IRC (Ping timeout: 480 seconds)
08:48:07 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
09:01:16 *** foodliker has quit IRC (Quit: User went offline on Discord a while ago)
09:44:58 *** reldred has quit IRC (Quit: User went offline on Discord a while ago)
10:49:33 <_glx_> Hmm GitHub annotations are supposed to be ignored
10:57:56 <_glx_> Interesting failure for #14337
11:34:50 *** WormnestAndroid has quit IRC (Remote host closed the connection)
11:35:09 *** WormnestAndroid has joined #openttd
11:46:34 <_glx_> Might be the new runner image
12:13:36 <_glx_> ok happens locally too after upgrading VS
12:13:36 *** moll has quit IRC (Read error: Connection reset by peer)
12:43:05 <_glx_> indeed one of the listed changes in new SDK mention Windows.UI.ViewManagement
12:52:07 <peter1138[d]> Drop Windows support when? 😉
12:53:41 <_glx_> seems they touched stuff in this area with newer sdk
13:10:05 <_zephyris> Steam-sync'd savegames would be kinda nice...
13:11:25 *** Smedles has joined #openttd
13:20:06 <_zephyris> Why not, last time I poked the documentation it looked relatively easy
13:20:44 <_zephyris> I vaguely ermember a debade on here about what should be sync'd, just savegames, settings, newgrfs, etc.
13:22:30 <truebrain> _zephyris: Easy to enable, but my only request has been that someone works out what should and shouldn't be sync'd. So far, nobody has put any effort in that, from what I can tell. Which is a shame 🙂
13:24:41 <truebrain> (last time this was a conversation, there were some strong opinions about what should (and shouldn't) be sync'd .. no clue if those strong opinions are still around 😛 But having a list helps out with not having to deal with those strong opinions every other week 😉 )
13:27:33 <truebrain> owh, and Steam warns (but doesn't prevent) storing files larger than 256MiB. They just note it might not be in an optimal location in terms of upload/download speed, relative to the user.
13:28:36 <_zephyris> Ah, so opinions vs. consensus, with no record of old discussion..
13:30:33 <truebrain> Most "opinions" when we released on Steam were like: "we should sync ALL files", without defining what "ALL" is. Others had strong opinions about content downloaded from BaNaNaS vs local. Crashlogs, autosaves ... screenshots, etc.
13:30:33 <truebrain> Honestly, it is much more about someone putting in the effort than anything else.
13:31:40 <truebrain> I still hope one day we can offer cloud saves in-game, but financially that has been a bit of a problem 😛 I tried last time Cloudflare offered to sponsor us, if they are willing to sponsor that part too. But they weren't at the moment in time 🙂 Already happy they pay for the rest 😄
13:32:18 <_zephyris> Cloudflare storage seems to be getting good
13:32:26 <_zephyris> Oh wow, zero egress fee storage
13:32:37 <_zephyris> Well, maybe better luck persuading them next time!
13:32:40 <truebrain> Cloudflare R2 has been amazing to us; it hosts BaNaNaS 😛
13:33:10 <truebrain> And we do know an average savegame is ~3MB, tnx to Survey! 🙂
13:34:15 <_zephyris> Well, I'm stuck on a plane soon, so I'll keep it in mind to think about!
13:35:06 <truebrain> 🛫 ✈️ ✈️ ✈️ ✈️ ✈️ ✈️ 🛬
13:39:44 <_zephyris> Thinking about the cloud, while up in the clouds...
14:39:47 <peter1138[d]> Integrate with NextCloud/OwnCloud/etc, so players can use their own storage 😉
14:40:38 <andythenorth> object categories aren't lexically sorted, I'm guessing that's so authors can control the order?
14:40:46 *** Wormnest has joined #openttd
14:47:59 <peter1138[d]> It's load-order.
14:50:36 <andythenorth> I'll reorder the src
14:50:51 <andythenorth> I tried to do it by using the name string, but that's so faff
14:51:43 <peter1138[d]> Don't forget about savegame compatibility.
15:20:55 <michi_cc> @glx Possible fix (to be verified with different SDKs):
15:20:55 <michi_cc> ```diff --git a/src/video/win32_v.cpp b/src/video/win32_v.cpp
15:20:55 <michi_cc> index 798a2dc475..aa541acc08 100644
15:20:55 <michi_cc> --- a/src/video/win32_v.cpp
15:20:55 <michi_cc> +++ b/src/video/win32_v.cpp
15:20:57 <michi_cc> #include <versionhelpers.h>
15:20:59 <michi_cc> #if defined(_MSC_VER) && defined(NTDDI_WIN10_RS4)
15:20:59 <michi_cc> #include <winrt/Windows.UI.ViewManagement.h>
15:21:01 <michi_cc> +#pragma comment(lib, "RuntimeObject")
15:21:03 <michi_cc> #include "../safeguards.h"
15:44:39 <michi_cc> Hmm, `RoGetActivationFactory` is listed as Windows 8+. So either the new SDK makes a nice mess of the promise of WinRT being a header-only/not hard-linked library, or we lost Windows 7 without knowing about it.
15:47:03 <_glx_> seems it's better to do `#pragma comment(lib, "windowsapp")`
15:48:51 <_glx_> any winRT is win8 minimum
15:52:31 <michi_cc> So "Windows 7 / 8 / 10 / 11" would be a lie now. Win 7 and 8 are totally out of support of course.
15:54:16 <_glx_> but we use winRT stuff only for 10+ (light/dark themes)
15:54:42 <michi_cc> Yes, but if it hard-links to these Ro... functions, it will still fail directly on load.
15:55:21 <michi_cc> Let's try to find out if they were also used with older SDK
15:56:56 <michi_cc> Okay, this is a change with the 10.0.26100.0 SDK version, and a breaking change as well. If I link and build with 10.0.22621.0, `RoOriginateLanguageException` and `RoGetActivationFactory` are *not* listed with DependencyWalker.
15:57:09 <_glx_> new runner image uses "Windows 11, Build 10.0.26100.4188 (released 5/2025)", previous was using "Windows 11, Build 10.0.26100.3916 (released 4/2025)"
16:04:12 <michi_cc> _glx_: Using windowsapp instead of RuntimeObject makes the EXE refer to even more API-MS-WIN-CORE-... dlls, so even worse 😦
16:04:38 <michi_cc> I wonder if a runtime DLL import stub could help...
16:08:33 <_glx_> oh 22621.0 does `impl::load_runtime_function(L"combase.dll", "RoGetActivationFactory", handler, fallback_RoGetActivationFactory)`
16:10:45 <_glx_> I guess the fallback was still in 10.0.26100.3916
16:16:09 <michi_cc> _glx_: Yeah, found that just now, too. So basically 26100 SDK is breaking all old Windows versions with this 😦
16:16:37 <_glx_> previous 26100 was fine
16:16:47 <michi_cc> Let's try to cheat by dong the runtime loading ourself.
16:34:16 <michi_cc> Okay, hacky but it apparently works. I'll prepare a PR.
16:36:50 *** Smedles has joined #openttd
16:54:49 <peter1138[d]> What about reverting the original change?
16:57:27 <_glx_> original change is a nice visual effect
16:59:08 *** digitalfox has joined #openttd
16:59:08 <digitalfox> That “nice visual effect” is awesome for anyone using dark mode on Windows (atleast in 11)
16:59:14 <michi_cc> Or we just de-list Windows 7, even the extended support was stopped in January 2023.
16:59:18 <_glx_> (yeah reverting it works, I tried)
16:59:45 <peter1138[d]> michi_cc: Does the survey tell if that is viable?
17:00:24 <peter1138[d]> > Windows 7 1.2%
17:01:37 <_glx_> michi_cc: January 2023 was end of support for 8.1
17:01:54 <michi_cc> And 2.7 (Win 7) + 1.2 (Win 8.0) for the 15 survery
17:02:54 <michi_cc> And now the PR run still failes due to the macos runner annotations
17:03:03 <peter1138[d]> Probably some older NewGRF developers "stuck" on Windows 7...
17:03:26 <_glx_> can't even run modern nml
17:04:05 <michi_cc> Well, the actual coding work is done either way. We can revisit if the SDK breaks it again, but until then the proposed change does not produce any angoing work.
17:04:23 <_glx_> yeah the .github messages are supposed to be ignored, but they aren't
17:07:02 <_glx_> ha it's not a warning, yeah for consistency
17:12:22 <belajalilija> so long as you guys continue windows 10 support for a few years im very happy))
17:20:13 <Rubidium> well, I guess it's mostly a question of how hard Microsoft starts unsupporting Windows 10
17:33:46 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
17:44:06 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:49:59 <_glx_> ok the annotation-check is in theory easy to fix, but npm is not cooperative
17:51:42 <_glx_> for some reason "build" changes almost everything in dist/index.js, and "test" fails
17:56:11 *** WormnestAndroid has joined #openttd
20:51:56 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:42:14 *** WormnestAndroid has joined #openttd
22:13:39 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:58:54 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
23:02:04 *** akimoto has joined #openttd
continue to next day ⏵
