Time Nick Message 01:24 crazylad while the latest commit added a highly anticipated feature, it looks like it may have caused some regression or something 01:25 crazylad whenever I go in the ContentDB menu and click on a package, all the screenshots are downloading and the whole hypertext scroll container glitches every formspec update 01:25 crazylad I already opened an issue -> #17313 01:25 ShadowBot https://github.com/luanti-org/luanti/issues/17313 -- Glitchy rendering of hypertext most notably in ContentDB menu (possible regression?) 04:51 Sheriff_U3 Testing is welcome! https://codeberg.org/luanti-tools/luanti-version-scanner 10:38 MinetestBot 02[git] 04AFCMS -> 03luanti-org/luanti: Migrate Docker CI to `docker/github-builder` (#17124) 13df04879 https://github.com/luanti-org/luanti/commit/df04879066de6eb94ca43996822a6dfacc74feca (152026-07-02T10:37:16Z) 10:53 cheapie Izzy: I only have one real data point for that, and that is that when I played it with an RX 550 plugged into a JH7110, the bus (PCIe 2.0 x1) was rather often the bottleneck 10:53 cheapie Although really that CPU struggled more than the bus did 11:05 Izzy I guess we will find out 11:15 Izzy oh cool a starfive 11:15 Izzy probably pretty comparable to the G5 11:19 cheapie I'm waiting for the PCIe stuff for the SpacemiT K3 to be upstreamed, once that all happens I'd like to try the same thing with that chip 11:19 cheapie That one has faster PCIe too, IIRC PCIe 3.0 x2 routed to that slot 11:24 cheapie FWIW "7z b" comes up with 4770 on the JH7110, while on the K3 running it on the X100 cores gives 16391 and on the A100 cores 9685 11:24 Izzy Looks like any AGP card in these machines will be limited to PCI-e 1.0 speeds, because AGP doesn't work on ... pretty much anything not x86 11:40 Izzy Huh. 7z b is neat. The G4 gets 511, the G5 732, and my desktop 65895 11:41 Izzy (desktop being a 5800X3D). My 4770 bench machine gets 22361 11:59 cheapie 142681 on my desktop here (5950X with a few other things running) 12:03 Izzy wew 12:03 Izzy guess it's not the X3D part making the difference there 12:13 cheapie Probably not, but it helps that the 5950X has twice as many cores and clocks a bit higher 13:01 Ingar and it hasn't exploded so far 13:02 sfan5 watch for the magic smoke 13:03 * cheapie emits magic smoke at sfan5 15:13 MTDiscord https://klipy.com/gifs/smoke-grenade-smoke 18:08 mazes_84 Hi I got a problem of encoding with my minetest/luanti client: ééé -> ééé 18:08 mazes_84 any hint as I compile myself, it happens only since ~5.15 18:09 MTDiscord you don't need to use both just luanti suffices also where are you finding this? 18:37 bgstack15 mazes_84: as a reality check, have you gone back to compile 5.14 with the exact same compiler and environment that produced the suspect client? Does the same-compiler-environment 5.14 build also suffer the same problem? 18:38 bgstack15 The next steps that I would see would be to then enforce various `LC_ALL=C.UTF-8` when running the client, to see if that makes a difference. 18:38 Krock I do not remember any changes related to encoding in the near past 18:39 bgstack15 (My advice is based on a GNU/Linux environment; I don't know about such encodings in a Windows environment) 18:39 Krock Windows also has an option to use the UTF-8 code page, though I don't think that's enabled by default 18:40 mazes_84 ok so there may be an error related to lua ? I've seen that lua-utf8 package was missing luajit support (hoped it was the issue but in vain) 18:40 Krock Luanti does not depend on lua-utf8. If you're using that, try dumping inputs and outputs to see where it starts to fail 18:42 mazes_84 it's also very long time I didn't encounter encoding issues 18:47 mazes_84 enforcing variables does not change anything, the weird point is that input is broken on all devices 18:47 mazes_84 I wonder i it may be related to SDL recent changes (keycodes against keystrokes I guess) 18:48 mazes_84 without being sure 18:49 mazes_84 hope it's a thing that will solve alone 18:50 mazes_84 the only version I can switch to is 5.11 (and mistake happened well after) 19:09 [MatrxMT] The keycode/scancode changes do not affect text encoding though (and is also unrelated to text input) 19:13 mazes_84 ok, thanks 19:16 mazes_84 anyway copy paste works as a charm, so encoding is ok. feels like my keyboard sends iso-8859 in this context 20:14 [MatrxMT] As said earlier: have you tried using LC_ALL=C.UTF-8 ? 20:23 mazes_84 y5nw: yes, I may wrote a bit much but I replied to this already