Implement fuzz testing about x64asm HOT 4 CLOSED

The current test infrastructure doesn't check memory operands. This is a pretty glaring omission. I think the longer-term solution here is to implement a fuzz-tester.

(I've marked this as low priority since STOKE essentially IS a fuzz tester that's been running for about a year now.)

from x64asm.

Well... I did it...

The results are... not as great as I expected... Guess that's what happens when you actually implement testing for real.

from x64asm.

If it makes you feel better, the fuzz testing on the validator tests the
sandbox and x64asm pretty hard too. But I've yet to encounter a single
sandbox/x64asm bug. You could even say that we've formally validated
the sandbox/assembler on some select inputs...

On 12/03/2014 12:53 AM, eric schkufza wrote:

Well... I did it...

The results are... not as great as I expected... Guess that's what happens when you actually implement testing for real.

---

Reply to this email directly or view it on GitHub:
#20 (comment)

from x64asm.

Yeah. I was thinking about that after I wrote this last night.

The errors that appear are superficial, I think. They have to do with
producing hex that's different from g++, but not necessarily incorrect.
Whatever results we're getting out of fuzz testing the validator are much
more confidence inspiring.

On Wed, Dec 3, 2014 at 1:37 AM, Berkeley Churchill <[email protected]

wrote:

If it makes you feel better, the fuzz testing on the validator tests the
sandbox and x64asm pretty hard too. But I've yet to encounter a single
sandbox/x64asm bug. You could even say that we've formally validated
the sandbox/assembler on some select inputs...

On 12/03/2014 12:53 AM, eric schkufza wrote:

Well... I did it...

The results are... not as great as I expected... Guess that's what
happens when you actually implement testing for real.

---

Reply to this email directly or view it on GitHub:
#20 (comment)

—
Reply to this email directly or view it on GitHub
#20 (comment).

from x64asm.