r178518 - [analyzer] Teach invalidateRegions that regions within LazyCompoundVal need to be invalidated

[Jordan Rose]

On Apr 1, 2013, at 18:28 , Anna Zaks <ganna at apple.com> wrote:

> +        // Note: the last argumet is false here because these are
> +        // non-top-level regions.

Typo: "agument".


This is fine for now, but I hope we can kill the three sets of regions...they're already a bit hard to follow.

- top-level non-const regions (passed to checkRegionChanges and checkPointerEscape)
- all non-const regions (passed to checkRegionChanges and checkPointerEscape)
- top-level const regions (passed to checkPointerEscape's const variant)

- all const regions (we currently don't model this but we might some day)

I guess we save a little bit of effort by caching the top-level regions in a list instead of extracting them from the CallEvent again, but I'm not sure it's worth the additional complexity in either notifyCheckersOfPointerEscape or checkRegionChanges.

Jordan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130401/d2e77c0a/attachment.html>