[cfe-dev] [analyzer] Random SVal hierarchy questions.

[Artem Dergachev via cfe-dev]

Hmm, i think i can answer myself regarding conjured structures, after a 
bit of thinking. There's really nothing wrong with symbols of structural 
type - in fact, it has been some kind of a dream for me to consider 
complicated constraints on structures, and especially on containers - 
consider a solver that can handle constraints like "this symbolic set is 
a subset of that symbolic set", or SymSymExpr's like "this set is an 
intersection of these two sets", etc., which can be achieved with 
structure-type symbols. Probably even strings would be modeled better 
with string-symbols and constraints than with stores. So i think it's 
all right to keep the current situation, even though we're not quite 
using these capabilities yet, and i withdraw this idea.