osudroid / osu-droid

http://osudroid.moe
Apache License 2.0
504 stars 75 forks source link

Save failed replay And Fixed a bug #102

Closed None1637 closed 3 years ago

None1637 commented 3 years ago

Can save failed replay Can save replay in offline Fixed a bug that star calculate no calculate DT/NC/speed modify move the pp display (score scene) to left-bottom

Rian8337 commented 3 years ago

The star rating calculation bug has actually been fixed, but it's up to you if you want to revert the change related to it or not. It should be fine this way as well.

Rian8337 commented 3 years ago

After testing in-game, I have two concerns:

Rian8337 commented 3 years ago

Just need answers for my questions.

None1637 commented 3 years ago

After testing in-game, I have two concerns:

* In replay, sliders occasionally gives 100 despite not being pressed. The video below should show it (bad audio, but the point is still there).
  https://user-images.githubusercontent.com/52914632/111894058-3103c900-8a3a-11eb-8995-6d478fe31a14.mp4

* Every hitobject that appears/should appear after the play fails is counted as miss, which is not something that we want? Unless this is required to tackle possible replay bug.

In replay, sliders occasionally gives 100 despite not being pressed. -This is a bug. I think that's why replay can get higher combo than real play. Maybe I can fix this bug. Every hitobject that appears/should appear after the play fails is counted as miss, which is not something that we want? Unless this is required to tackle possible replay bug. -If don't counted as miss, Every hitobject that appears/should appear after the play fails will be hit automatically in the replay. Count them as miss is the easiest way to solve this problem.

Rian8337 commented 3 years ago

This is a bug. I think that's why replay can get higher combo than real play. Maybe I can fix this bug.

Isn't that the original replay bug, which is caused by movement emulation? This one seems different since no cursor is present.

If don't counted as miss, Every hitobject that appears/should appear after the play fails will be hit automatically in the replay. Count them as miss is the easiest way to solve this problem.

The play would have failed already, so it shouldn't matter? I reckon this had something to do with the bug we've discussed earlier.

None1637 commented 3 years ago

Isn't that the original replay bug, which is caused by movement emulation? This one seems different since no cursor is present.

the original replay bug Include this bug.

The play would have failed already, so it shouldn't matter? I reckon this had something to do with the bug we've discussed earlier.

It does have something to do with bugs, and If don't count as miss, the replay would not fail.

NeroYuki commented 3 years ago

i assume this can show off the bug better https://www.youtube.com/watch?v=0lfl53hofEQ (skip to 0:15)

None1637 commented 3 years ago

i assume this can show off the bug better https://www.youtube.com/watch?v=0lfl53hofEQ (skip to 0:15)

this bug is fixed in my previous commit