[PATCH] D35450: [analyzer] Support generating and reasoning over more symbolic constraint types

[George Karpenkov via Phabricator via cfe-commits]

george.karpenkov added a comment.

> I can also split out the APSInt fix into a separate patch.

Yes please.

In general I really dislike arbitrary "complexity cutoffs", as they make further debugging hard and may lead to weird bugs: could you explain why is that required and can not be avoided?


https://reviews.llvm.org/D35450