Search results

I did stumble upon this when looking up where the conditionals could go. Wondered if it meant that what I wanted to add was already available, but quickly understood that it required you to be the one constructing the Request, which here, was not always the case... Care to point me to...

Tests-wise, there's currently no way to debug outgoing GenericRequests. Any method that ends with / has logic that centers around a call to RequestThread.postRequest is therefore pretty much un-testable. So, I was wondering if there would be any demand for something which would try to help with...

The code looks up choice adventures by ID, so I sorted the choice adventures by ID.

Well, that was kind of what some of the latest discussion was about: how else? They can't be sorted by name because of conflicts And if the sorting is done arbitrarily, you'd need to make sure one choice adventure can't fit in more than one "category" at once, otherwise someone will look into...

This is more than wild brainstorming; this is extremely close, if not a complete match, with what I went with: Excerpt: package net.sourceforge.kolmafia.persistence.choiceadventures; public class ChoiceAdventureDatabase { private static final Map<Integer, ChoiceAdventure> database = new...

This is the very downside I was preoccupied with regarding the "split big changes into small steps" approaches: the individual step's values are scrutinized without the knowledge of how they help/fit into a broader change. This rework is just meant to be a start. The goal is to end up being...

For 1), Hee^3 mentioned remembering that the unwieldiness of ChoiceManager is a shared feeling, though no citation/quote was added (not that it would've been easy, given that it must have been from hearing it here and there over time...) The re-arrangement he mentions is the one happening in...

I guess a question I need to ask is: have you actually taken a look at the patch, or are you solely basing that comment off of the recent exchanges? (this isn't a "look before you talk" kind of critique. This is a "I would like to know what information you had access to, in order to rule out...

But can you really compare the two? RuntimeLibrary is a collection of arbitrarily defined functions. This means that they can be arbitrarily broken up. They don't have an "order" other than alphabetically, possibly. Their existence isn't bound to anything, so we don't have to worry about one...

Can't the super/abstract class's constructor(s) simply do the registration themselves? Like... could just become abstract class SuperClass { SuperClass() { ChoiceBuilder.registerChoice( this ); } } class SubClass extends SuperClass { }

The creation of new choice adventures, mostly. It's not "imperative", it's just "a nice benefit".

If we had 1 choice adventure per file/class, how should each file/class be named? By its choice adventure name? By its choice adventure number? Each class still needs to be instantiated. This means that if we have a whole class dedicated to a choice adventure, it still requires a static scope...

Not sure whether "What's the goal" refers to splitting them along multiple files, or the "by choice adventure" refactor altogether, so... Reasons to split them: 1- Very big files cause editors to do things like mess up the formatting whenever a change is added. 2- In pull requests, GitHub can...

All right, this indeed seems feasible; insistence dropping... So we're back to: how should they be organized? Organizing them solely by zone may not be the best option. They are a lot, and I mean a *lot* (but not, like, "half of them" a lot), of choice adventures with an undefined location, as...

Any insight into how we should? I would really like to insist that whichever way we use should preserve the order of the choice adventures (i.e. avoiding "putting all choice adventures related to the elemental charters in this file, putting all choice adventures related to dreadsylvania in that...

So... should it really just be one giant file..?

Isn't it already how things are right now? ChoiceManager is 20k lines long, so you're going to have to grep (or any other text-finding tool) either way, afaik. This could be... an interesting challenge. Would it help if I were to manage to enforce this?

Bump. This seems to have been mostly ignored in the past two months. In the meantime, suggestions went from patches to git PRs. I can turn this into a PR, but knowing if there is any objection/concern to the concept would still be appreciated. Should I also still go method-by-method? (i.e. a...

May I get a heads up on the situation? Is it that we're waiting for the transition to github to continue, or is it that A) you're busy with something else (dealing with bugs being more important than new features), or B) there's a problem with the patches/something missing? (or C), something I...

Update: pruned the "modify the signature of CompositeType.getDataType( Object )" part of the patch. I'll look into that issue an other time.