Times are UTC Toggle Colours
00:56:03 *** Flygon has quit IRC 01:38:34 *** Wormnest has quit IRC 01:48:08 *** tokai|noir has joined #openttd 01:48:08 *** ChanServ sets mode: +v tokai|noir 01:54:59 *** tokai has quit IRC 02:00:02 *** keikoz has joined #openttd 06:10:06 <locosage> TileX, TileY 06:11:06 <locosage> z in coming from height map but there are several functions for that too 06:12:27 <tramrider> Thankyou! I shouldn't need Z in this scenario. I'm just working within roadstop tile(s). 06:33:51 <DorpsGek> [OpenTTD/OpenTTD] Kuhnovic approved pull request #12285: Fix #12280: Allow ships 90 degree turns if only choice https://github.com/OpenTTD/OpenTTD/pull/12285#pullrequestreview-1935789754 06:44:27 <truebrain> dwfreed: truebrain at openttd dot org will do fine. Tnx π 06:44:37 <dwfreed> ack 06:45:06 <dwfreed> You've got the whole /64, right? even though you use like an /80 for the discord bridge 06:45:15 <truebrain> Yup 06:45:39 <dwfreed> ack, will adjust in a few minutes 06:45:46 <truebrain> β€οΈ 06:47:45 *** Ox7C5 has joined #openttd 07:03:23 <dwfreed> truebrain: you now have effectively unlimited connections 07:04:06 <truebrain> Awesome, tnx π that should prevent people from not being able to speak π 07:04:14 <dwfreed> heh 07:04:56 <truebrain> Will restart the bridge in a sec, as my bot won't reconnect the puppets in its own π 07:07:47 *** lapingvino has quit IRC 07:07:47 *** merni has quit IRC 07:07:47 *** coil has quit IRC 07:07:47 *** rutoks has quit IRC 07:07:47 *** klote has quit IRC 07:07:47 *** jfs has quit IRC 07:07:47 *** ogi300 has quit IRC 07:07:47 *** ajmiles has quit IRC 07:07:47 *** emperorjake has quit IRC 07:07:47 *** bendn has quit IRC 07:07:47 *** tony_pixel has quit IRC 07:07:47 *** kamnet has quit IRC 07:07:47 *** dwfreed[d] has quit IRC 07:07:47 *** nanapipirara has quit IRC 07:07:47 *** lea001 has quit IRC 07:07:47 *** brickblock19280 has quit IRC 07:07:47 *** frosch123 has quit IRC 07:07:47 *** _jgr_ has quit IRC 07:07:47 *** kuhnovic has quit IRC 07:07:47 *** xarick has quit IRC 07:07:47 *** belajalilija has quit IRC 07:07:47 *** locosage has quit IRC 07:07:47 *** peter1138 has quit IRC 07:07:47 *** _glx_ has quit IRC 07:07:47 *** DorpsGek_vi has quit IRC 07:07:47 *** teifion has quit IRC 07:07:47 *** tycekcz has quit IRC 07:07:47 *** truebrain has quit IRC 07:07:47 *** ahyangyi has quit IRC 07:07:52 *** DorpsGek_vi has joined #openttd 07:08:44 *** truebrain has joined #openttd 07:08:44 <truebrain> That is a long list ... seems thr Discord is getting more and more popular π 14:01:38 *** nielsm has joined #openttd 14:12:49 *** talltyler has joined #openttd 14:12:49 <talltyler> michi_cc: π I donβt know if this motivates you, but if/when you finish that consists patch, I have a fleshed-out push-pull/trains-can-back-up plan that I should be able to implement pretty easily π 14:17:33 <merni> peter1138: Oh man, Indian scripts are even better 14:19:57 <merni> https://www.unicode.org/notes/tn10/ 14:33:32 <michi_cc> talltyler: Why do you think I started the consist branch in the first place π But yes, I have too many much too unfinished branches sitting around π 14:33:51 <merni> merni: or https://www.unicode.org/versions/latest/ch12.pdf which is impressively detailed 14:34:02 <merni> lots of things in there that I took for granted 14:39:43 <talltyler> Well, whenever you get around to it, I will happily do my part and hopefully make your work worth it beyond simply cleaning up tech debt π 14:40:19 *** Ox7C5 has joined #openttd 14:58:52 <peter1138> Well 15:38:03 *** Smedles has quit IRC 15:38:12 *** Smedles has joined #openttd 15:58:56 <truebrain> michi_cc: that branch is scary well created; the commits are very easy to read π 16:01:24 <truebrain> and I am out of ideas how to make benchmarking work on GitHub .. the ubuntu runners have all different CPUs, so comparing there is a bit pointless .. and on the M1s I can only use the CPU time, which just fluctuates too much. Meh. 16:01:27 <andythenorth> consists eh π 16:02:44 *** Wormnest has joined #openttd 16:03:59 <peter1138> My idea was just to use `std::unique_ptr<BaseConsist>` inside Vehicle. IMHO pools only really exist because we didn't have managed containers/pointers back in the day. 16:04:14 <peter1138> Though I guess a consist pool makes it easy to iterate just consists. 16:06:45 <andythenorth> are consists consistent? 16:07:00 <peter1138> There's also more things that can in the consist information, e.g. current_order 16:07:14 <peter1138> and unitnumber. 16:11:01 <truebrain> Makes a lot of places more clear, instead of the endless ->Front() π 16:19:53 <truebrain> peter1138: Should at least be shared_ptr not? 16:20:58 <peter1138> I didn't go all that far, my aim was just to reduce the amount of memory used by vehicles that don't need any of this information. So only the front vehicle has the unique_ptr. 16:21:10 <peter1138> But yes, michi's work is superior. 16:21:18 <truebrain> Ah, like that, gotcha 16:21:27 <peter1138> I was just looking after you commented about BaseConsist being made larger by somebody π 16:21:47 <truebrain> I wouldn't dare! π 16:23:12 <peter1138> You weren't complaining, just mentioned it! 16:24:28 <LordAro> truebrain: dedicated VPS! 16:24:33 <LordAro> (for benchmarking) 16:25:05 <truebrain> Meh. More infra to support π¦ 16:25:29 <truebrain> MacOS has kperf, but implementing that without a Mac is challenging 16:25:43 <truebrain> There is no CLI, just a headerfile 16:33:19 <peter1138> https://cdn.discordapp.com/attachments/1008473233844097104/1217873206782722128/image.png?ex=66059bcf&is=65f326cf&hm=5b48559fd8f902a6da168655f749c9ddfe1829921a84cd5d04948c3ac8c48a93& 16:33:19 <peter1138> Hmm, better? 16:33:49 <peter1138> Left: master. The number is the group index, so shared across all companies. 16:34:00 <peter1138> Right: pathc. The number is a per-company group number. 16:34:16 <peter1138> (Ab)Using FreeUnitIDGenerator. 16:34:26 <truebrain> Wait, it was global? Clearly I play too much single player ... 16:34:34 <truebrain> Better: yes 16:34:39 <LordAro> definitely less confusing 16:34:42 <peter1138> Also, the first group you create, in single player, is "Group 0" 16:35:52 <peter1138> https://discord.com/channels/142724111502802944/477434889508093952/1217866689967030292 inspired me and about 20 minutes later... 16:37:07 <peter1138> As it uses the same scheme as unit numbers, if you buy out a company the group numbers will change too. 16:37:29 <peter1138> I imagine most people rename groups anyway, but it still bugged me π 16:40:34 <truebrain> So another branch in the endless sea of your branches? π 16:42:19 <_glx_> don't forget all the patch/diff not in branches π 16:47:59 *** gelignite has joined #openttd 16:48:25 <peter1138> All those random stashes π 16:57:21 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #12296: Codechange: split initiating of joining and identification of the client https://github.com/OpenTTD/OpenTTD/pull/12296 17:02:44 <peter1138> > 27 files changed, 6 insertions(+), 405 deletions(-) 17:02:46 <peter1138> A good day coding... 17:15:54 <DorpsGek> [OpenTTD/OpenTTD] PeterN opened pull request #12297: Change: Add per-company group numbers. https://github.com/OpenTTD/OpenTTD/pull/12297 17:17:19 <peter1138> I knew it. 17:18:55 *** gelignite has quit IRC 17:20:41 <DorpsGek> [OpenTTD/OpenTTD] PeterN updated pull request #12297: Change: Add per-company group numbers. https://github.com/OpenTTD/OpenTTD/pull/12297 17:21:08 <peter1138> Is it cheating to update the regression test result? 17:23:55 <DorpsGek> [OpenTTD/website] 2TallTyler updated pull request #307: Add: Blog post about timekeeping in OpenTTD 14 https://github.com/OpenTTD/website/pull/307 17:24:43 <DorpsGek> [OpenTTD/website] 2TallTyler updated pull request #307: Add: Blog post about timekeeping in OpenTTD 14 https://github.com/OpenTTD/website/pull/307 17:26:28 <peter1138> ctest-before-push, pom te pom. 17:29:20 <andythenorth> maybe I should try this again https://www.scenario.com/ 17:29:33 <andythenorth> although last time I tried to train a model it made literal noise, and I got bored 17:41:57 <_glx_> peter1138: if the change is expected 18:34:23 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #12296: Codechange: split initiating of joining and identification of the client https://github.com/OpenTTD/OpenTTD/pull/12296#issuecomment-1998085484 18:36:28 <truebrain> _glx_: did you happen to test if the vcpkg fix also unbreaks our nightly? 18:45:52 <_zephyris> Generic image synthesis neural networks are useless for pixel art... 18:54:46 <truebrain> You sound surprised π 18:55:37 <truebrain> truebrain: tested myself, just to be sure; and indeed it fixes the issue. So now they just have to merge. 19:01:24 <_glx_> just tested in my docker 19:02:20 <_glx_> hmm icu fails for me 19:03:24 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #12296: Codechange: split initiating of joining and identification of the client https://github.com/OpenTTD/OpenTTD/pull/12296#issuecomment-1998158409 19:03:44 <truebrain> worked for me; could even bake a binary 19:06:09 <_glx_> ```sh-4.4# cat /vcpkg/buildtrees/icu/config-x64-linux-dbg-err.log 19:06:09 <_glx_> configure: WARNING: unrecognized options: --disable-silent-rules 19:06:09 <_glx_> ./../src/c-74_1-src-8ca76704fa.clean/source/configure: line 4620: syntax error near unexpected token `0.20' 19:06:09 <_glx_> ./../src/c-74_1-src-8ca76704fa.clean/source/configure: line 4620: `PKG_PROG_PKG_CONFIG(0.20)'``` 19:06:35 <_glx_> but dinner time 19:15:01 <truebrain> enjoy! 19:27:14 <andythenorth> _zephyris: I'm not sure why 19:27:22 <andythenorth> they should be able to learn it trivially 19:27:33 <andythenorth> all my brain is doing is repeating patterns 19:27:58 <andythenorth> and we've both written different kinds of pixel art generators with primitive (ish) programming approaches 19:28:28 <andythenorth> I guess the problem is knowing whether the patterns mean anything, or are just junk shapes 19:31:18 <xarick> hi 19:59:49 *** greeter has quit IRC 20:01:02 *** greeter has joined #openttd 20:05:19 <andythenorth> https://cdn.discordapp.com/attachments/1008473233844097104/1217926556857532477/image.png?ex=6605cd7f&is=65f3587f&hm=c9319c2c1b0cc59061c2dcb159f07310bb6ee7f1f01815903ce04788df69460a& 20:05:19 <andythenorth> silly FIRS 20:05:24 <andythenorth> goes it throw out GS? 20:16:06 <michi_cc> truebrain: Did you think of a teaser yet? 20:16:20 <truebrain> I still have a day π 20:28:22 <andythenorth> https://cdn.discordapp.com/attachments/1008473233844097104/1217932357605003284/image.png?ex=6605d2e6&is=65f35de6&hm=957adca1056e6a617fa0951507a9697c4f36031221256d61aae438d3d84d0382& 20:28:22 <andythenorth> should I add 'choose a random engine' from these variants? 20:28:35 <andythenorth> hmm, no I'll probably run out of varact2 IDs π 20:30:53 <reldred> andythenorth: YES. OH GOD YES. 20:32:41 <andythenorth> kind of wonder if the game should do 'random from variants' 20:32:53 <andythenorth> oh we could do it as a callback, and filter the list π 20:33:58 <andythenorth> so it would be an "engine" in the buy menu, with an ID, but it actually builds another ID, with a pointer to the "engine" ID that originally built it 20:34:03 <andythenorth> (for autoreplace reasons) 20:34:14 <peter1138> Open a discussion on github. 20:34:17 <andythenorth> oh yes 20:34:18 <peter1138> Make a proposal. 20:34:19 <andythenorth> we could 20:34:27 <andythenorth> it has worked before 20:34:44 <peter1138> HouseSpec 20:34:56 <andythenorth> currently playing OpenTTD though 20:34:59 <andythenorth> killing cacodemons 20:41:05 <DorpsGek> [OpenTTD/OpenTTD] PeterN opened pull request #12298: Codechange: Shuffle members to reduce vehicle size. https://github.com/OpenTTD/OpenTTD/pull/12298 20:43:19 <_glx_> truebrain: I forgot to apply a "cp" 20:43:27 <truebrain> DOH 20:43:37 <_glx_> aclocal stuff is important π 20:45:31 <_zephyris> andythenorth: I suspect its foundational in how the networks are designed... The input format/resolution, the network architecture, the output are all designed around high resolution images. Every output wants smooth pixel-pixel transitions! You'd want to start from scratch. (I've had related problems at work) 20:45:36 <_glx_> with properly configured image it works 20:48:18 <_glx_> it built, everything's fine 20:49:18 <_glx_> I missed <https://github.com/OpenTTD/OpenTTD/blob/master/.github/workflows/release-linux.yml#L62> when pasting into the docker terminal 20:49:22 <DorpsGek> [OpenTTD/OpenTTD] andythenorth started discussion #12299: Spec for player-buildable 'randomised' vehicles. https://github.com/OpenTTD/OpenTTD/discussions/12299 20:50:11 <andythenorth> _zephyris: It seems that the vector space for patterns of pixels, in 8bpp, would be relatively low dimension, compared to photo-real images 20:50:16 <andythenorth> (to me at least) 20:50:57 <andythenorth> and also we know there are clear x / y / z orientations in the sprite, which means we might be able to infer faces of the object 20:52:00 <andythenorth> so there might be a relatively small set of vectors corresponding to 'building walls' 20:52:12 <andythenorth> yada yada 20:52:48 <andythenorth> TBH, my real-world experience was with n-dimensional vector hyperspaces for text analysis, which predated LLMs 20:59:00 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #12300: Feature: encrypted game sockets https://github.com/OpenTTD/OpenTTD/pull/12300 22:45:15 <_glx_> (openttd cmake) 22:49:42 <salut3585> do you mean /baseset/openttd ? 22:50:12 <LordAro> PATH as in environment variable 22:50:21 <LordAro> or just set the variables 22:50:41 <_glx_> they are cmake variables 22:50:59 <LordAro> not related to openttd checkout 22:51:13 <LordAro> just needs to be somewhere such that running `grfcodec` actually works 22:51:18 <LordAro> (or set the cmake variables) 22:54:07 <salut3585> I will check it. Thank you! 22:58:38 <_glx_> if you can run grfcodec, grfid and nforenum from anywhere, then openttd cmake should find them 23:12:54 *** HerzogDeXtEr has quit IRC 23:14:42 <salut3585> I read that nforenum was added into grfcodec 23:15:11 <LordAro> the repos were merged, they're still separate executables 23:15:20 <LordAro> a very long time ago 23:24:01 <salut3585> do you have a link on githab NFORenum 23:24:20 <salut3585> it does not work https://dev.openttdcoop.org/projects/nforenum 23:24:47 <LordAro> https://github.com/OpenTTD/grfcodec 23:33:26 <salut3585> ok, what do I need to add? 23:33:39 <salut3585> export NFORENUM_EXECUTABLE=/usr/local/nforenum 23:33:39 <salut3585> export GRFID_EXECUTABLE=/usr/local/grfid 23:33:39 <salut3585> export GRFCODEC_EXECUTABLE=/usr/local/grfcodec 23:34:06 <LordAro> they need to be cmake variables, not environment variables 23:34:18 <LordAro> so cmake -DNFORENUM_... 23:34:31 <LordAro> or, if you add your symlinks into /usr/local/bin instead, it'll probably all just work 23:34:39 <LordAro> as /usr/local/bin will likely be on PATH 23:37:12 <salut3585> LordAro: it has not worked I will try to use the make variables 23:49:10 <_zephyris> https://cdn.discordapp.com/attachments/1008473233844097104/1217982889581609040/image.png?ex=660601f5&is=65f38cf5&hm=92ca9bb26cbcbb9f01bc35cb9d9d778cb11f2dc2affac893279d5d9d8ce44735& 23:49:10 <_zephyris> Stylistically this is quite far from standard Hebrew... But I think it matches the serif font style quite well. 23:51:27 <_zephyris> https://cdn.discordapp.com/attachments/1008473233844097104/1217983464176091146/image.png?ex=6606027e&is=65f38d7e&hm=80626a60cd6f32b21b0499d2a69f5c6b8843c1c49a4dc20cc1cf437315008d0f& 23:51:27 <_zephyris> I think the sans is is a safe design.