None input maps to "pass" instead of DefaultBattleOrder

[cameronangliss]

There is no way to send pass in poke-env. Having DefaultBattleOrder as the default option for the orders in DoubleBattleOrder is misleading, and it really should be pass. The user could just provide DefaultBattleOrder to the DoubleBattleOrder constructor if they want one of the sides to be default, so this is a strict empowerment for the user.

[hsahovic]

This is a good change.
Can you add some unit tests?

[codecov]

Codecov Report

Attention: Patch coverage is 86.66667% with 2 lines in your changes missing coverage. Please review.

Project coverage is 85.10%. Comparing base (f458350) to head (62095ac).
Report is 128 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #659      +/-   ##
==========================================
+ Coverage   83.38%   85.10%   +1.72%     
==========================================
  Files          39       42       +3     
  Lines        3918     4404     +486     
==========================================
+ Hits         3267     3748     +481     
- Misses        651      656       +5     

[cameronangliss]

@hsahovic this is being blocked by bugs in poke-env not being able to handle Zoroark properly😞 all other tests pass. I guess the good news is that this PR exposes a bug, but the bad news is that Zoroark is a nightmare. @caymansimpson if you're interested in taking down one of the most infamous problems of all time, you're invited hahaha

[caymansimpson]

haha I don't think it's worth our time to tackle Illusion; it involves keeping multiple possible battle states and integrating speed/move inferences into the core of poke-env, which if comprehensively done, will require 10K+ lines of code (my inferences work is 95% accurate at 6K lines; there are so many niche interactions).

[hsahovic]

why is this change in this PR?

[cameronangliss]

So, this was a little while ago, but if I remember right, this was to try and simplify the code and fix Zoroark bugs on our side of the battle specifically (since that's all the request tells us about, is our own side).

[cameronangliss]

to be clear, it's a replacement for the code being removed a few lines further below.

[cameronangliss]

I definitely could test taking it out real quick and see what happens, whether that breaks stuff.

[cameronangliss]

Alright, I remember it all now. So, yes, this is a necessary fix, and it is for Zoroark. This is also just a much easier code chunk to look at than what it's replacing. Basically, the request will always put the active pokemon in the first and second positions in the pokemon list, and although you could be worried that we aren't correctly filtering out fainted pokemon, if you look at the _get_active_pokemon method in double_battle.py (which is used to give the active_pokemon property), you'll see that it is filtered there, so we were actually double-filtering before.

[cameronangliss]

I've now moved this change to #666, as that is where it is more relevant.

[hsahovic]

I think this accidentally got commits from #666?