[PATCH] D71224: [analyzer] Escape symbols stored into specific region after a conservative evalcall.

Tue Dec 10 14:18:59 PST 2019

xazax.hun added a comment.

In D71224#1778303 <https://reviews.llvm.org/D71224#1778303>, @NoQ wrote:

> In D71224#1778284 <https://reviews.llvm.org/D71224#1778284>, @xazax.hun wrote:
>
> > So I was wondering if we got the default right. Maybe a checker should do more work to get the escaping rather than more work preventing it?
>
>
> But that's exactly how it works right now(?) If you don't define `checkPointerEscape` you get no escaping, if you do extra work of defining it you get the exact amount of escaping that you want.

So basically what I am wonder/worrying about is the following:
The analyzer core will decide that the stack region is escaped and the checkers has no word about this. And from that time on the checkers have to do extra work each time there is a store or conservative call to find out if this escape corresponds to a region that was escaped earlier unwillingly (from the checker's point of view).

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71224/new/

https://reviews.llvm.org/D71224