@@ -39,6 +39,8 @@ PN may have issues with test 2 and 8 since the all offsets inference can't be us
Also note I did modify the source code a bit for PN. There were some checks provide by Wilander to make sure a test is still valid. These tests used a propolice dummy variable. In some situations these checks still make sense, but in light of variable reordering, in some cases this check is meaningless. A check is still needed to make sure the attack is still possible, so I commented out these checks when necessary and placed my own check in.
Note the flags in the make file. Wilander makes some indirect function calls. Normally MEDS will change these instructions, and PN will not see the call. I have turned off as many MEDS options as possible including this one so that a canary check is inserted before these calls.
