Issue with Action Grant/Remove

[Sirryan2002]

Whenever a Species changes it will call on_species_loss and on_species_gain which is also called in that species DM file.

When ever on_species_lost is called for something like slimes, it remove the actions recolor and regrow limbs

/datum/species/slime/on_species_loss(mob/living/carbon/human/H)
	..()
	if(grow)
		grow.Remove(H)
	if(recolor)
		recolor.Remove(H)
	UnregisterSignal(H, COMSIG_HUMAN_UPDATE_DNA)

However if you transform back into that species it will call on_species_gain and readd those actions

/datum/species/slime/on_species_gain(mob/living/carbon/human/H)
	..()
	grow = new()
	grow.Grant(H)
	recolor = new()
	recolor.Grant(H)
	RegisterSignal(H, COMSIG_HUMAN_UPDATE_DNA, /datum/species/slime/./proc/blend)
	blend(H)

So at this point, the mob has both the grow and recolor actions(and the icons), however when the mob transforms back into a different species again, on_species_loss is once again called, HOWEVER, it fails the if checks for both actions even though they were granted back to the mob.
both if(grow) and if(recolor) return false and therefor the actions are not removed from the player

I checked the variables and it appears that the mob loses the action /datum/action/item_action/button when switching around. If anyone could help me figure out why recolor and grow return false please share your wisdom.

[SteelSlayer]

When transforming as a changeling, we end up here:

/mob/living/carbon/human/proc/change_dna(datum/dna/new_dna, include_species_change = FALSE, keep_flavor_text = FALSE)
	if(include_species_change)
		set_species(new_dna.species.type, retain_damage = TRUE)
	dna = new_dna.Clone()
	[...]

The set_species() call here works fine, just as you suspect, the slime gets their grow and recolor because on_species_gain() was called. However the issue happens inside the Clone() proc, specifically the line: new_dna.species = new species.type

/datum/dna/proc/Clone()
	var/datum/dna/new_dna = new()
	new_dna.unique_enzymes = unique_enzymes
	new_dna.struc_enzymes_original = struc_enzymes_original // will make clone's SE the same as the original, do we want this?
	new_dna.blood_type = blood_type
	new_dna.real_name = real_name
	new_dna.species = new species.type // fix me
	[...]

That line creates a freshly new'd species datum which does not have any of the code from on_species_gain(), because that proc is never ran.

When we finally return to the change_dna() proc, dna = new_dna.Clone() resets our species to a blank species datum, erasing the grow and recolor (or similar) variables.

I suspect #15085 was the PR which created this bug.

[Sirryan2002]

Alright, this definitely explains it. I'll probably have to called on_species_gain again somewhere to fix this.