Joe Ranweiler
Joe Ranweiler
While adding X87 instructions, I've found that we don't model (or sometimes test) various input classes for X87 semantics. This surfaced when I started adding more exhaustive test inputs for...
## Summary Remill models some 80-bit x87 FPU registers with `f64`'s. Per discussion with @pgoodman, this was a design choice to prioritize bitcode portability. 1. What are the possible impacts...
This will let us implement `klee_is_symbolic(uintptr_t)`.
This relationship should be inverted. The current implementation was just convenient when implementing assertions that should trigger failures. See the commit message for d81cbba13788b262e2d10a8da35ca88121b22d9f.
Consider: - `-n` or `-j` for `--num_workers` - `-o` for `--output_test_dir`
Besides reading input from a directory as described in #3, how else can we enhance the native test harness? What is actually worth implementing? Ideas: - Report discrepant runs of...
We define the executor `-h` text in `common.py`, in the `parse_args()` class method. It is [hardcoded](https://github.com/trailofbits/deepstate/blob/master/bin/deepstate/common.py#L114) to `Angr` (which I think is technically mis-stylized, since it is an [anti-proper ](https://docs.angr.io/docs/faq.html#how-should-angr-be-stylized)noun)....
See: https://github.com/reem/rust-traitobject/issues/7 A maintained fork can be found at [`destructured_traitobject`](https://github.com/philip-peterson/destructure_traitobject#author).
Unless I missed something in the docs, this doesn't seem possible right now. This would pair naturally with the existing `data()` accessors that exist, which provide access to the (unparsed)...
Right now, theft provides `theft_random_double()`. This returns a `double` in the closed unit interval `[0, 1]`. What would be nice is some other floating point `alloc` builtins that do things...