Entities lists loaded with Entity:loadList now remember which files were used

Warden's Call changes

Add proc chance to Warden's Call

Previously, Warden's Call would always create a clone after a
successful attack. This was somewhat overpowered, but more importantly
it exacerbated the memory leak issue with temporal clones.

Now, Warden's Call activation is chance-based, capping at 53% chance at
5/5 talent level with 1.3 mastery. To compensate for the reduced
activation rate, the damage penalty on the clones has been reduced.

This should (hopefully) reduce the skill's performance degradation by
roughly half. When the memory leak is fixed, we can re-evaluate whether
the activation chance is still needed or if it should be increased.
However, I think this is probably better for game balance anyway. The
skill in its original form was stronger than other talents such as
Greater Weapon Focus, which is already considered strong by the
community.

This could be pushed back to v1.3.2 with the other cloning changes.

Misc

Who broke the Nightmare?

Previously, Warden's Call would always create a clone after a
successful attack. This was somewhat overpowered, but more importantly
it exacerbated the memory leak issue with temporal clones.

Now, Warden's Call activation is chance-based, capping at 53% chance at
5/5 talent level with 1.3 mastery. To compensate for the reduced
activation rate, the damage penalty on the clones has been reduced.

This should (hopefully) reduce the skill's performance degradation by
roughly half. When the memory leak is fixed, we can re-evaluate whether
the activation chance is still needed or if it should be increased.
However, I think this is probably better for game balance anyway. The
skill in its original form was stronger than other talents such as
Greater Weapon Focus, which is already considered strong by the
community.

This could be pushed back to v1.3.2 with the other cloning changes.

Merge branch 'cloning_fixes' of http://git.net-core.org/Effigy/t-engine4 into Effigy/t-engine4-cloning_fixes

Added a cleanup function for Arrow Stitching, since it spawns clones
the same way as Warden's Call. The cleanup function is the same.

Changed makeParadoxClone() to not copy fov and distance_map tables from
the source actor. Instead, we call ActorFOV.init() on the clone.

Updated a few comments for clarity.

Merge branch 'wardens_call_fix' of http://git.net-core.org/Effigy/t-engine4 into Effigy/t-engine4-wardens_call_fix

A few different things

Hi DarkGod, I baked you some code! :D

Contains 5 different things, which are separate should you not want all of them.
1: Changes disease spreading code to be more general to works nicely with any future or addon diseases.
2: Adds bleeding damage to Bone spells. They are a little bit weak and then I won't have to hack them for Witherer. :)
3: Balances some creatures with innate 100 points per turn resource regeneration.
4: Adds more zone auras which include handy tooltips for explaining things like finding bubbles when underwater. Also adds some flavourful modifiers.
5: Addresses a bug with the hasShield() change I found while testing this.

Small fixes

Fix 1: Improve efficiency of Arcane Combat info

Changed Arcane Combat's description generator to use table.concat, to
avoid making temporary string objects that need garbage collection.

Fix 2: Fix various typos

1. Since the customizable table cloning function may be useful for
other purposes than just making Chronomancy clones, I moved it to
engine/class.lua and renamed it cloneCustom(). It works like
cloneFull(), but allows specifying a table of keys/values to skip or
replace in the clone table.

If desired, I could instead just modify cloneFull() to have an optional
parameter for the "alt_nodes" table, but I didn't want to mess with it
without permission. There could be a conditional statement to pick
between two different recursive functions so the optional param doesn't
impact performance, if that's a concern.

2. Updated comments for cloneFull() and related functions for clarity.
The comment originally said it returns the clone and the number of
objects cloned, but it actually only returns the clone. I think the
comment may have been copied from somewhere else and not updated.

3. Renamed the "t" param from cloneFull() to "post_copy" to be more
descriptive and to be consistent with cloneCustom(). This has no
functional impact.

4. Significantly expanded the number of properties which are skipped
when making Chronomancy clones of an actor. Hopefully this will improve
performance and maybe even help with the memory leak. I didn't see any
errors during brief testing, but it probably merits more testing.

Changed Arcane Combat's description generator to use table.concat, to
avoid making temporary string objects that need garbage collection.

Adds more zone-wide effects in. About half of them explain special level mechanics, the rest are just for flavour or some minor gameplay spice.

I just remembered why I originally designed it this way: We only want
to skip if neither 'k' nor 'v' is specified in alt_nodes since we want
to allow substituting user-specified values for either.