Tavern Brawls in Fireplace
This has been a fun couple of weeks with commits from two new contributors! I’m very happy about the generally positive feedback on the quality of Fireplace. I even wrote some contributing guidelines for the occasion.
Since Tavern Brawls have come out, I have been focusing on separating rules from the Game object itself. 14f7e61 begins the separation by creating a BaseGame object and removing The Coin (random start) and Mulligan from it.
The default Game object is now declared as
Game(MulliganRules, CoinRules, BaseGame).
This is one of the rare times where mixins actually shine.
It’s very elegant, but still not ideal. I would much prefer a system where I can declaratively write rules for the Game object, just like I write them for regular Cards. Brawls could reuse such a system as well.
The first three Brawls are now implemented:
None of them was particularly hard once the mechanics for them were in place.
Banana Brawl was the trickiest of the bunch, requiring a slight
rework of death preparation logic
to allow for subclassing.
In addition, I had to
implement caching in
so that I could simply implement the “random banana” logic as a
class RandomBanana(RandomCardGenerator): cards = ("EX1_014t", "TB_006", "TB_007", "TB_008")
cards is a property on RandomCardGenerator, I could just override it with a static list.
It’s worth noting that in the real Banana Brawl, the picks are probably weighted.
This would still be very easy to do, by overriding the
pick() method which simply calls
in the default implementation.
As a sidenote, the Wild Magic Hero Power was the original reason why I started working on Shuffle and Copy. Even though they don’t look related at first glance, down the line, they led to the implementation of recursive actions which I used for Unstable Portal, and now Wild Magic.
Week 4 is currently running. Near-unanimously acclaimed as the most fun Brawl yet. I’m excited to implement
it in Fireplace, though I am running up against the issue of intercepting
Play actions. I’ll figure it out.
In unrelated news, the codebase has been moved to
under_score-style variable/method naming, to align with
PEP8. While I don’t agree with all its recommendations, naming
style is something I believe should align as much as possible with the official language style.
That move accounts for a huge part of the diff below.