Both trackered for 5.16.0
> --- In email@example.com, "Barak" <barak@> wrote:
> > The difference? One is an arcane caster and one is a divine caster.
> > Now, as far as I'm concerned the spell chooser should pay no
> attention to
> > the *characters* caster type... it's only use is to pick spells.
> The bane of every user: Working as designed/architected.
> You've hit a very subtle, but in my opinion important, point in making
> the SPELLS chooser "feature complete".
> The brackets define a context; from the docs: "The LEVELMAX sub-tag
> references all spells of the specified level and lower _within the
> context of the 'y' variable preceding the brackets_" (emphasis added)
> Means "List all of the arcane spells from level 0 to the *maximum
> castable arcane spell level* minus 1". Thus, it's paying attention to
> the caster's type, because you told it to: you told it to evaluate the
> MAXCASTABLE variable within the context of SPELLTYPE=Arcane.
> Thus for a divine caster, this is 0-1 = -1, so you'll never get any
> spells in the chooser (and can't take the feat).
> This is important to define the context, since this is the only way to
> produce the appropriate subsets based on the restrictions (otherwise
> you could never get things only on a specific class list, etc.). This
> example from the docs:
> is a classic example of why context is important when you consider a
> 15th level Cleric; 2nd level Sorcerer ... you WANT the context
> evaluation if you need to choose a castable Sorcerer spell for a feat.
> The good part is that the syntax should still allow you to produce the
> result you want (since it's *supposed* to be feature complete, but
> there's a bug, see below)
> Since the comma is effectively an AND function, you should be able
> "List any spell that is both
> - up to my maximum castable (of any spell) level minus 1
> - SPELLTYPE=Arcane
> [TM] DOC FREQ: Can someone get the ANY...SPELLTYPE example above into
> the docs, please, perhaps with the negative example From Barak's post
> to help explain context?
> Now, having said that, there IS a bug.
> [TM] CODE BUG: ANY happens to not be implemented as a "y" item (ugh),
> so that's a problem and makes it so you can't do what you want at this