Proper time advancement for gardening

[lapislosh]

Currently the way gardening works is when the player enters their mog house, the server will check if their plants need to be updated and if so, advances them to the next stage (or withers them). When the player examines their plants, a bit is flipped in the database to signify that it was examined for the current stage, and the bit is unset when the stage advances.

This has a number of problems like:

1. Stages only advance when you enter your mog house, meaning plants take longer to grow than they theoretically should
2. Players with multiple characters have their plants get out of sync with each other (typically the primary character advances faster due to more mog house entrances)
3. You can ignore a plant for several days and not have it wither as long as you don't enter your mog house to advance the stage
4. If a stage is shorter than the withering timeout, it's impossible for the plant to wither even if you never check it (unless you let it go for 7+ days or whenever the hard-coded max is)

I feel like this should be pretty easy to fix into something that makes more sense - the goal being that plants advance at a constant pace regardless of when you enter your mog house and withering happens if you don't check your plants once per ~24 hours. I don't know if any real gardening data exists to know exactly how it should work, but the current system is entirely fabricated so updating it to something that "feels" better seems reasonable.

There's two main problems causing the current state.

1. When a plant advances stages, it uses the current timestamp. Instead, it should use the previous timestamp + the previous stage length. Once a plant advances, it should try to keep advancing stages until it reaches a time past the current time.
2. The previous change would completely break the withering system, as plants would regularly be advancing multiple stages at once and always wither. A timestamp should be stored on the flowerpot any time the plant is examined, and that should be used as the time check for withering instead of the start of the current stage. From what I can tell, there are currently 4 unused bytes at the end of flowerpot data (0x14 - 0x17) which would fit this timestamp perfectly.

Before I do any work here (I think it's probably only about 10-20 lines of code total), I'd be curious to know if anyone has thoughts about why this might be a bad idea, what changes might be needed to ensure the PR isn't rejected, or any reasons why those extra flowerpot bytes can't be used.

[jmcmorris]

I'm not sure what gave you the idea that the current system is entirely fabricated but I can assure you it is not. I worked with someone (Wiggo iirc) on gathering retail captures for several weeks and then implemented what we have now. I believe the underlying system is retail accurate. I will admit that the plant data (seeds + crystal) results are somewhat fabricated as there was never much data accumulated for them as it'd take a very long time to do so.

You are welcome to investigate and see if I did happen to miss something. This project was like the second thing I did for the codebase and it was about 8 years ago at this point. If you are going to look into it I recommend digging into the pot extra data first as I believe that is where all the details live. Good luck!

[lapislosh]

Sorry, I didn't mean to imply it had no basis at all despite the words I used. The gardening results and the gardening strength system seem based in partial reality but it has never seemed like a true retail accuracy thing (due to the relative impossibility of figuring it out), so I have been under the assumption it was mostly educated guessing to get it kind-of close, and I further assumed the way stages advanced was also somewhat guessed and perhaps a technology constraint.

If that's not actually the case and the way plants advance over time is truly how it's supposed to work then my question here would be answered and there's no work to be done. I don't really have the means to try and prove any retail accuracy myself, so it was more of an inquiry about a potential fix that I would have guessed may have been more accurate than the current system (but maybe not).

[jmcmorris]

It's all good I know that a great deal of what was added to legacy DSP code was more fabricated than retail accurate. It's been a long time but I believe everything was tested for the gardening system including things like it not updating the plants unless you zone in.