More comprehensively check whether a GEP edge can induce a GEP object#1870
More comprehensively check whether a GEP edge can induce a GEP object#1870mbarbar wants to merge 5 commits into
Conversation
| // If we aren't treating the first field as equivalent to the base, no matter the | ||
| // offset, we may induce a GEP object. If the field index isn't 0, no matter how we | ||
| // treat the first field (0), we may induce a GEP object. | ||
| return !Options::FirstFieldEqBase() || 0 != normalGepCGEdge->getConstantFieldIdx(); |
There was a problem hiding this comment.
Could we do the opposite way for the condition to make it intuitive?
If(Options::FirstFieldEqBase() && 0==normalGepCGEdge->getConstantFieldIdx())
Return false
There was a problem hiding this comment.
Done, if you could take a look again.
|
Would be good to also include your test case (rational for this pr) in https://github.com/SVF-tools/Test-Suite/tree/master/src/basic_c_tests |
I'm looking at the 0 AP offset and believe the test has far larger issues. I think we can add the test after that is sorted. I believe it to be a coincidence that this issue showed up from that test case in the sense that if we didn't have the AP offset problem, this ff-eq-base stuff wouldn't have happened. |
|
nander: ander: Diff: |


No description provided.