search

ppy / osu

rhythm is just a *click* away!
https://osu.ppy.sh
MIT License
15.06k stars 2.23k forks source link

Mod "Wind Up" and "Wind Down" reset speed rate to initial on results screen #17264

Open frenzibyte opened 2 years ago

frenzibyte commented 2 years ago

Noticable specifically on beatmaps with tracks that are longer than the beatmaps themselves (e.g. https://osu.ppy.sh/beatmapsets/184#osu/8707).

https://user-images.githubusercontent.com/22781491/158472635-18b96c7f-2988-4b26-8317-b864b6b1f302.mp4

Should stay in final rate similar to how it behaves in song select, rather than abruptly switch to initial (and weirdly enough, fix itself when clicking on the results score card).

Discussed in https://github.com/ppy/osu/discussions/17262

Originally posted by **jkl2k2** March 15, 2022 Sorry for the wordy title. If a beatmap's song continues to play in the results screen while using "wind up" or "wind down," whether because the song has an outro or the map is unfinished, then the rate (and pitch if enabled) of the song will not match the final rate from wind up and wind down. The song's rate will only be corrected after clicking the results card to view the expanded breakdown on the right. However, mods like double time and nightcore will continue to apply the correct rate/pitch without having to click the card. I am on version 2022.314.0-lazer Logs: [database.log](https://github.com/ppy/osu/files/8256390/database.log) [network.log](https://github.com/ppy/osu/files/8256391/network.log) [performance.log](https://github.com/ppy/osu/files/8256392/performance.log) [runtime.log](https://github.com/ppy/osu/files/8256393/runtime.log) [updater.log](https://github.com/ppy/osu/files/8256394/updater.log)
bdach commented 2 years ago

@frenzibyte can you actually reproduce this on master? because I can repro on 2022.314.0 but not on master. was in the middle of bisecting this, in fact.

frenzibyte commented 2 years ago

Turns out I was a couple of commits behind master and 2022.314.0 (at https://github.com/ppy/osu/commit/dc1e5a456a6ec7a29ed79f77dfcd6bc875001a7a), I suppose this is an issue that has been "resolved" in current master then...

bdach commented 2 years ago

bisection turned up 4a3e3aba65a1cece80e164fefc7622f853637d8c which looks obviously bogus. and i managed to not get the bug once on 2022.314.0, so i'm banking on this being some sort of race that was made less frequent by the recent player optimisations.