Re: [PATCH] Accept plists when serializing and parsing JSON

[João Távora]

Stefan Monnier <address@hidden> writes:

> Notice the use of "change" in that docstring.  Looking up the value of
> a dynamically-scoped variable doesn't *change* global state, but it
> *uses* global state.

Yeah, but it's the "change" bit that usually counts: it's (sneakily)
writing shared state that has the bug potential alluded to by Yuri.

That said, lisp is particularly bad territory for purity advocates. Many
many functions use global state, and many more change it just by consing
up something, or interning a symbol.