NaN present in JSON

[TakingFire]

Current behavior

Some older beatmaps produce a message containing NaN in the Performance object. In this example, multiplierProgress:

{"status":{"performance":{"currentSongTime":10,"relativeScore":0.886956512928009,"multiplier":0,"combo":4,"maxCombo":4,"currentMaxScore":0,"lastNoteScore":0,"hitNotes":4,"missedNotes":0,"rank":"S","passedBombs":0,"softFailed":false,"rawScore":0,"batteryEnergy":null,"passedNotes":4,"score":0,"hitBombs":0,"multiplierProgress":NaN,"energy":0.539999961853027}},"time":1662254394851,"event":"scoreChanged"}

This is invalid JSON and can cause issues with parsing.

Expected behavior

NaN is replaced with null or 0 to keep the JSON valid.

Reproduction steps

1. Start an old map (for example: https://beatsaver.com/maps/45)
2. Catch any socket message containing the Performance object
3. Attempt to parse JSON

Environment

BeatSaber version: 1.24.1 HTTP Status version: 8.0.2 Platform: SteamVR, Windows

[denpadokei]

Thank you for your report. Start investigating.

[denpadokei]

The problem did not reproduce. Could you upload the log when the phenomenon occurred? (InstDir)/Logs

[TakingFire]

On further testing, it only occurs when viewing Replays. I believe it may be old replays, rather than old levels. Here is the _latest.log, it contains the beginning of two songs that both cause the issue.

Here is also my replay file for https://beatsaver.com/maps/45, if it is useful: 76561198257862880-Give It Up-9BCFBD06C6CEAC3E4FF44093FBC33AB83587108B.zip

[denpadokei]

It seems to be by design that multiplierProgress returns incorrect values with old replay data. Make HTTPSiraStatus always return 0 instead of NaN.

Could you please test this DLL and confirm that 0 is returned? HttpSiraStatus-8.0.3-rc1-bs1.21.0-b844f89.zip

[TakingFire]

I still appear to be getting NaN for multiplierProgress. Demonstration video:

https://user-images.githubusercontent.com/38340654/188304737-69d63d9c-e3dd-4d5f-8be3-3b4ebb824117.mp4

[denpadokei]

HttpSiraStatus-8.0.3-rc2-bs1.21.0-074797c.zip Here is the output of the performance as a log. Could you upload the output log after playing the NaN replay here?

[TakingFire]

This version works perfectly, here's the log showing it is 0 now. I'm not sure why the other .dll didn't work.

Many of the replays were giving invalid JSON, so I'm glad to see it working properly now! It's somewhat strange that score and currentMaxScore are mostly 0, but it's likely just due to the outdated replay.

[denpadokei]

thank you for making sure I release this version.

Since the implementation of the replay function is a black box, it is difficult for me to deal with it any further. You'll probably need to contact the ScoreSaber team, but your proposal will probably be rejected because the results and effort you're getting aren't worth it.