Speed. For doing max int32 scans, this will be significantly faster than python (and also easier to parallelize).
On my machine scanning the Traveling Cart stock:
- python (single core): 21,000 seeds per second (28 hours for 2^31-1)
- C# (single core): 280,000 seeds per second (~13x faster, 2-2.5 hours for 2^31-1)
- C# (parallel): 1,800,000 seeds per second (~85x faster, 20 minutes for 2^31-1)
Remix bundle scanning:
- python (single core): 16,000 seeds per second (37 hours for 2^31-1)
- c# (single core): 315,000 seeds per second (~20x faster, 2 hours for 2^31-1)
- c# (parallel): 1,300,000 seeds per second (~80x faster, 28 minutes for 2^31-1)
When you need to do more complicated searches, that speed difference will be even more pronounced.
- Remixed Bundle scanning (for details see Remixed Bundles)
- Traveling Cart scanning (for details see Traveling Cart)
- Location forage (docs in progress)
NOTE: this uses .NET 5 to be in line with Stardew Valley. .NET 5 is not supported anymore/you'll get compilation warnings but it'll still work assuming you have the .NET 5 SDK installed.
The SMAPI docs have some great instructions for setting up a C# development environment here. Assuming you have things set up, you should be able to clone this repo and start setting up your searches.
** NOTE:** I can't seem to get the post-build command to correctly copy the data folder into the bin folder cross-platform. After you build, you'll need to manually copy the data folder into the bin folder (under either the bin/Debug/net5.0/ or bin/Release/net5.0/ folder that is generated).
- Add docs for location forage
