Developer Notes - 2021-02 Ruleset Updates

[Trenloe]

On a technical aside, when 360+ lines of code are involved, concerns are not separated. Certainly each case is something that SW would have to evaluate priorities for. As the code is currently set up, it is extremely difficult for an extension developer to make improvements without overriding the method entirely (e.g. to support lifestealing attacks in 5e).

Serious question - where would you want integration points within the current code?

As an example, the CoreRPG code has various points where custom functions can be inserted. Which points would you see as being useful to add integration points for extension developers? And which ruleset/s are you referring to?

[SilentRuin]

If people are expecting pristine code in a project that has been written by so many people (plus the ruleset authors and the extension authors) - with the resources involved here - dream on. This will always be a work in progress IMHO. I'm sure others disagree but that is my opinion on having worked on "many hands stirring" type of projects with 3rd party libraries tossed in for chuckles Getting it working > making it look pretty.

[celestian]

Serious question - where would you want integration points within the current code?

As an example, the CoreRPG code has various points where custom functions can be inserted. Which points would you see as being useful to add integration points for extension developers? And which ruleset/s are you referring to?

I was going to sit down and type out the various hooks I'd find useful but the more I thought about it the more I determined I needed to override the entire thing to do what I'd need.

I think there could be some work to consolidate (even move to API) Effects tho. Getting it out of the LUA environment might help improve it's performance. I'd really like to see something done to improve performance on large chunks of data (iterating over large number of tokens, effects or npc is ponderous. How that can be improved, I dunno.

[SilentRuin]

... I determined I needed to override the entire thing to do what I'd need.

SO hate when that happens. Why I call templates evil when trying to override their .lua especially.

[deer_buster]

On a technical aside, when 360+ lines of code are involved, concerns are not separated. Certainly each case is something that SW would have to evaluate priorities for. As the code is currently set up, it is extremely difficult for an extension developer to make improvements without overriding the method entirely (e.g. to support lifestealing attacks in 5e). In that regard, the code cleanup itself does basically nothing from a feature perspective, rather it is a force multiplier empowering the community to make optional enhancements more efficiently and with more stability. This in turn can actually accelerate feature development, as it is not terribly uncommon for reliable extensions to be integrated into the rulesets. Even if community contributions aren't considered, in my experience code cleanup on an active codebase has paid for itself within 6 months 100% of the time, so then its mostly a matter of which 6 month window makes sense for the roadmap if one is to consider the damage management portion of the ruleset an active codebase.

Sounds like you just volunteered. Good luck breaking it down to useful sections that can have good integration hooks with each other and extension writers. I'd be very interested in finding out how long it takes to get just one ruleset's applyDamage code fully optimized.

[deer_buster]

I was going to sit down and type out the various hooks I'd find useful but the more I thought about it the more I determined I needed to override the entire thing to do what I'd need.

I think there could be some work to consolidate (even move to API) Effects tho. Getting it out of the LUA environment might help improve it's performance. I'd really like to see something done to improve performance on large chunks of data (iterating over large number of tokens, effects or npc is ponderous. How that can be improved, I dunno.

I thought someone had mentioned that an effects engine overhaul was in the cards...maybe dreaming that though...

[deer_buster]

Serious question - where would you want integration points within the current code?

As an example, the CoreRPG code has various points where custom functions can be inserted. Which points would you see as being useful to add integration points for extension developers? And which ruleset/s are you referring to?

I think he was specifically referring to 5e...

[Moon Wizard]

@bmos,
You're correct on ActorManager.isPC(). I fixed in the notes.

Thanks,
JPG

[celestian]

SO hate when that happens. Why I call templates evil when trying to override their .lua especially.

I've got so much "extra" maint work during updates like this because of 1 thing I need to tweak and it's not a "global" script I can just override a function for ;(

Not being able to merge/replace when using versioning on controls (version=">=4") makes me cry a little also.

[MeAndUnique]

Sounds like you just volunteered. Good luck breaking it down to useful sections that can have good integration hooks with each other and extension writers. I'd be very interested in finding out how long it takes to get just one ruleset's applyDamage code fully optimized.

That's a reasonable interpretation; I would happily make a merge request if there were a mechanism to do so. Roughly 2 hours of work between last night and this morning to accomplish a reasonable first pass.

Serious question - where would you want integration points within the current code?

As an example, the CoreRPG code has various points where custom functions can be inserted. Which points would you see as being useful to add integration points for extension developers? And which ruleset/s are you referring to?

The 5E ruleset is my primary interest, though I'm sure the concept would translate to others. I'm less interested in explicit integration points, though having one at the end would be useful, and more interested in more fine-grained definition of the functionality. There's a rather tight relationship between the Single Responsibility Principle and the Open and Closed Principle. In other words, if the code is broken down so that each function is responsible for doing one thing well, being able to extend any one thing becomes a very straight-forward process.

See here.