This commit adds Chi nodes to the successor relation and accounts for
them in the CFG, but does not add them to the SSA data graph. Chi nodes
are inserted for partial writes to any VirtualVariable, regardless of
whether the partial write reaches any uses.
This enables the addition of new instructions in later phases of IR
construction; in particular, aliasing write instructions and inference
instructions.
This adds a `NewOrNewArrayExpr.getPlacementPointer` predicate and uses
it in `Alloc.qll` to detect when a `new`-expression is not an
allocation.
User-defined replacements for `operator new` may not be allocations
either, but the code continues to assume that they are. It's possible
that we want to change this assumption in the future or leave it up to
individual queries to decide on which side to err. It's hard to
statically tell whether `operator new` has been overloaded in a
particular file because it can be overloaded by a definition that is not
in scope but is only linked together with that file.
An slightly invalid AST can cause IR construction to generate extremely bad IR. This change provides a single place to detect invalid ASTs, and to skip IR construction for the affected functions.
This change provides a mechanism by which a query can tell the IR package to only create IR for certain functions. This is mostly useful for "PrintIR.qll", which uses this feature to avoid the expense of creating IR for functions that aren't going to be printed.
This commit inserts an `Uninitialized` instruction to "initialize" a local variable when that variable is initialized with an initializer list. This ensures that there is always a definition of the whole variable before any read or write to part of that variable.
This change appears in a different form in @rdmarsh2's Chi node PR, but I needed to refactor the initialization code anyway to handle ConditionDeclExpr.
