Toshi Piazza

Some things that are TODO: - [ ] Unicorn integration - [ ] Support multiple states in merge - [ ] Support no common ancestor in merge - [ ]...

travis failure test_strtol passes for me locally and windows failures seem unrelated

This also makes it easier to deobfuscate OLLVM's opaque predicates, which we currently don't bother doing.

@egrimley Hmm that's odd, I tried on my raspberry pi (so ARM, not AArch64) and everything seemed to work fine. I'll take a closer look to make sure it really...

I'll do more research on the ARM failure (I actually didn't modify `test_asm_123` in my commit so I'm wondering what's up with that). As for the AArch64 failure, I think...

I think I have the Raspberry Pi 2 model, which is running Debian Jessie.

Interesting: I chrooted using debootstrap to pull debian stretch, and I was able to replicate the failure in debian stretch -- natively. The `drx_buf-test.c` program itself fails natively even without...

Looks like the problem was that on Debain Stretch, `$r4` was assumed not clobbered, but in Debian Jessie it was restored correctly. The pthread handling code immediately dereferenced `$r4` after...

I guess it's unclear if this is expected or desired to work, since the docs for Medium::Ip suggest that Medium::Ip is an abstraction designed for TUN devices... If this is...

This may be harder to fix than how I initially thought. When sending a packet on a socket created with AF_PACKET + SOCK_DGRAM it looks like you need to specify...