Luanti logo

IRC log for #luanti, 2026-11-19

| Channels | #luanti index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:03 TheWild Today I just discovered Starfox Adventures has been decompiled to the point it has been ported to PC, and yeah, it's available for Linux too.
00:05 mrkubax10 joined #luanti
00:08 Swift110-mobile joined #luanti
00:09 TheWild Last time I played it was over 10 years ago. On Dolphin Emulator. Despite getting 30-40% of performance, the game got me hooked so bad I wanted to buy actual GameCube and actual disc. But didn't happen - I played up to "snowmobile race" then moved on to other things.
00:13 TheWild Now I can't really engage in games anymore because there's no wonder, no "yay, I made to the next level" and I switch into autopilot mode after few minutes of playing. :/
00:26 pgimeno cheapie: can't stay but implemented BLAKE2s-256 and it is faster than sha256sum.S (about 4s execution time in singleplayer with default RVC params). There are two problems: one is that all loops are unrolled, making it about 10kb. The other is that it doesn't output the correct result, but I plan to debug and fix it. I don't think it will be noticeably slower after fixing.
00:27 cheapie OK, so marginally faster then (the sha256 implementation in passworddoor was optimized a bit and now takes about 5 seconds)
00:28 pgimeno optimized wrt a previous version of the same file?
00:28 cheapie This one has optimizations: https://cheapiesystems.com/git/rvcontroller/tree/assembly/passworddoor/sha256sum.S
00:28 cheapie As opposed to this one: https://cheapiesystems.com/git/rvcontroller/tree/assembly/sha256/sha256.S
00:29 pgimeno ah right, I compared it to sha256sum.S, not to sha256.S
00:29 pgimeno btw, sha256.S doesn't print the right result for me, not sure what I did wrong
00:29 cheapie This is the commit that did it: https://cheapiesystems.com/git/rvcontroller/commit/assembly/passworddoor/sha256sum.S?id=f212e8e0d9f86148ba216bcc19639b77fbb18c4c
00:29 cheapie sha256.S printed the right result with all of the strings I tested, although I of course didn't test every possible string
00:30 cheapie Keep in mind that it does _not_ add a trailing \n at the end of the string, so if you're making the file with a text editor then that \n will change the result
00:30 wsor4035 Hey Luanti modding devs: the devex team is looking for you, *yes you*, to tell us what would make your life easier making mods/games/etc. vote at https://www.poll-maker.com/p/3kov3mnk9s and/or give some feedback in https://irc.luanti.org/luanti-devex/
00:30 wsor4035 <shill alert> check out the current to tools luma - cdb cli https://github.com/luanti-org/luma and flint - format/lint/converter https://github.com/luanti-org/flint <shill alert>
00:32 pgimeno cheapie: I just try the empty string for starters, which is mentioned in wikipedia with the correct value (I also compare it with sha256sum.S which does print the right result)
00:32 pgimeno gotta fly now
00:33 cheapie Huh, it's giving me wrong answers too now. It used to not, no clue what changed
00:35 cheapie pgimeno: Fixed it, it was leftover testing stuff causing it to append trailing garbage to very short inputs
00:36 cheapie Longer inputs (at least 2 characters) were working OK
00:56 mrcheese joined #luanti
01:21 chilledfrogs joined #luanti
01:51 ndorik-zenith joined #luanti
02:47 SwissalpS joined #luanti
03:00 SwissalpS joined #luanti

| Channels | #luanti index | Today | | Google Search | Plaintext