486c/rosu-memory

Lightweight, cross-platform memory reader for osu! but in rust

rosu-memory's Issues

Reading loop occasionally resets

Loop of finding osu folder -> reading static signatures -> actual reading loop seems to break on Windows machines. Requires platform-specific investigation, as I was unable to replicate the issue even once

Example output:

Seems to be related to #44 in some form, as I've not seen anyone replicate only one of the two so far

Adjust beatmap-related calculations to use Gradual calculations

rosu-pp's GradualDifficultyAttributes and GradualPerformanceAttributes allow for higher efficiency in calculations. However, to use it you need to track object deltas. Issue is: reading loop is called every 300ms. In some cases there might be no new objects. process_next_n_objects() doesn't allow for delta = 0, so we need to adjust for that slightly

pp is 0 on results screen and song select

Should show the pp for the score when clicking on a score on the leaderboards or after coming from gameplay.

For song select it should probably just show the SS pp for the selected beatmap and mods.

Cannot access memory at negative address in windows

This error have occured when I ran 1.0.0 release build on windows

fix: check ruleset_addr for zero's

rosu-memory/src/reading_loop.rs

Line 326 in 250ad37

let ruleset_addr = p.read_i32(

<- if this value is zero then we can't read anything

Takes a minute or so for the overlay to start

As of the latest commits, it now takes slightly longer for the key overlays to begin working.

calculates wrong star rating for converted maps

afaik, gosu-pp supports converted beatmaps, however when i put on a converted beatmap, i notice that the calculated star rating is incorrect. i'm assuming the pp value is also wrong, but i haven't checked.

`new_map` seems to never actually be true

Simple

        if new_map {
            println!("new map!");
        }

never fired even ONCE across minutes of changing maps, game states etc
requires investigating

Failure after "random beatmap" is spammed

Steps to reproduce:

• Open osu and rosu-memory
• Open song select screen
• Spam F2

Result:
Shows an error of bad address read

Expected outcome:
No errors

Reach data parity with StreamCompanion

Currently rosu-memory lacks data compared to SC

Useful data points:

• BPM and current BPM (partly added in #19)
• username (added based on replays in #19)
• current score (added in #19)
• current letter grade (added in #28)
• background image of the current beatmap
• current skin name (added in #26)
• map ranked status (added in #27)
• current hp (added in #19)
• map star rating

Allow to specify pid manually

Converted sr not calculated correctly for converted catch maps

Noticed in the process of investigating #52, it seems that catch sr and probably pp isn't calculated correctly at all, and it only gives the osu!standard sr.

Prepare for 1.0 release

• #20
• #30
• #17
• Benchmark reading loop & static addresses loop and document result's into README
• Document changes about osu! folder auto-detection
• Port counters that use map background
• Port counters that have map progress in it

"data.current_pp" seems to display higher pp than actually obtained

This bug is kinda hard to talk about since I can't really replicate it with replays. This has happened most notably when I've full combo'd maps, and noticed that the pp counter is off by like 50-150pp. Just today when I got 500~, it said 630~ which was very odd.

I don't think it will be trivial for me to get more information regarding this issue 🙁.

Current (gradual) calculations occasionally panic

Steps to replicate:

• Run osu! and rosu-memory
• Watch a replay (seems to be non-specific)

Expectation:
rosu-memory runs just fine

Result:
rosu-memory panics with an error on L596

Seems to be a regression from my initial fix of potential extra triggers of gradual.nth(). Requires investigating

I wasn't able to replicate this behaviour on linux so far, might be related to a different issue

"data.gameplay.slider_breaks" always returns 0

Playing on Arch Linux.

Replicate:

1. Start rosu and osu!
2. Play a map and sliderbreak
3. Check console.log for an overlay

don't hang whole reading loop when pp failed to calculate

"data.fc_pp" doesn't show actual "if fc" pp

I'm assuming that the purpose of data.fc_pp is to display the pp value for an fc if you didn't sliderbreak/miss. However this doesn't seem to be the case. It seems like it (almost) displays the value if you got an fc INCLUDING the misses, which is obviously not possible. I'll upload pictures showcasing the problem and the replay.

The value displayed on the key overlay:

The pp for "full combo" with the misses:

The actual pp for full combo:

"Skipped gameplay reading, probably it's not ready yet"

I can't get it to work on Arch Linux running the program as sudo ./linux_rosu-memory --osu-path /home/flyslime/.local/share/osu-wine/osu\!, auto detection doesn't change anything. Sometimes it doesn't print the message Skipped gameplay reading, probably it's not ready yet, but most of the time when I go into gameplay it does. None of the overlays work, neither in song select or in-game.

I notice nothing happens if I start the program while osu! is running, however when I open osu! afterwards we get the following log:

To replicate:

1. Open the rust program
2. Open osu! (takes 5-10 seconds)
3. Play a map

Logs:

Spawned server!
Process not found!

Location:
    src/main.rs:97:34
sUsing provided osu! folder path
Reading static signatures...
Cannot found signature F8 01 74 04 83 65

Location:
    src/structs.rs:197:19
Using provided osu! folder path
Reading static signatures...
Cannot found signature F8 01 74 04 83 65

Location:
    src/structs.rs:197:19
Using provided osu! folder path
Reading static signatures...
Cannot found signature F8 01 74 04 83 65

Location:
    src/structs.rs:197:19
Using provided osu! folder path
Reading static signatures...
Starting reading loop
[src/main.rs:185] values.gameplay.passed_objects = 0
[src/main.rs:186] values.playtime = 3958
[src/main.rs:187] values.prev_state = PreSongSelect
[src/main.rs:188] values.state = SongSelect
Trying to read bad address
Address: 4, Length: 4


Location:
    src/reading_loop.rs:242:30
Trying to read bad address
Address: 10, Length: 4


Location:
    src/reading_loop.rs:107:10
Skipped gameplay reading, probably it's not ready yet

Score data isn't reset when changing state to ScoreSelect

Gameplay data (300/100/50/0, combo, etc) doesn't get reset upon entering song select screen, only when starting a new beatmap