Commit Graph

12 Commits

Author SHA1 Message Date
Dave Bartolomeo
b5a3edfdae C++: FunctionIR -> IRFunction 2019-03-12 11:28:22 -07:00
Robert Marsh
17ad124c9e C++: remove VariableAddress from points_to test 2019-03-07 13:14:56 -08:00
Robert Marsh
7e30ce0c09 C++: add phi node support to escape analysis 2019-03-07 13:14:56 -08:00
Robert Marsh
97c11a5222 C++: points-to for argument-returning calls 2019-03-07 13:14:55 -08:00
Robert Marsh
c70bd285de C++: assume arguments to virtual functions escape 2019-03-07 13:14:49 -08:00
Robert Marsh
2c94a8887d C++: test for virtual functions in escape analysis 2019-03-07 13:14:49 -08:00
Robert Marsh
6089172554 C++: escape analysis for this parameters 2019-03-07 13:14:49 -08:00
Robert Marsh
466e110338 C++: add new interprocedural escape analysis 2019-03-07 13:14:48 -08:00
Robert Marsh
bd39698528 C++: test changes for interproc escape analysis 2019-03-07 13:14:48 -08:00
Dave Bartolomeo
4c23ad100e C++: Rename a few IR APIs
There are a few IR APIs that we've found to be confusingly named. This PR renames them to be more consistent within the IR and with the AST API:

`Instruction.getFunction` -> `Instruction.getEnclosingFunction`: This was especially confusing when you'd call `FunctionAddressInstruction.getFunction` to get the function whose address was taken, and wound up with the enclosing function instead.

`Instruction.getXXXOperand` -> `Instruction.getXXX`. Now that `Operand` is an exposed type, we want a way to get a specific `Operand` of an `Instruction`, but more often we want to get the definition instruction of that operand. Now, the pattern is that `getXXXOperand` returns the `Operand`, and `getXXX` is equivalent to `getXXXOperand().getDefinitionInstruction()`.

`Operand.getInstruction` -> `Operand.getUseInstruction`: More consistent with the existing `Operand.getDefinitionInstruction` predicate.
2019-02-06 22:43:49 -08:00
Dave Bartolomeo
aa267c8302 C++: Force LF for .c,.cpp,.h,.hpp 2018-09-23 16:23:52 -07:00
Dave Bartolomeo
4086a8909b C++: Fix a couple IR-related tests to handle new directory tree
Also moved those tests under the IR test directory, so I'm less likely to forget them next time.
2018-09-04 09:05:33 -07:00