[PATCH] D158707: [analyzer] Fix a few size-type signedness inconsistency related to DynamicExtent

[Ding Fei via Phabricator via cfe-commits]

danix800 added inline comments.


================
Comment at: clang/lib/StaticAnalyzer/Checkers/MPI-Checker/MPIChecker.cpp:180-183
+      SVal CountReached =
+          SVB.evalBinOp(State, BO_GE, Idx, Count, ASTCtx.BoolTy);
+      if (!CountReached.isUndef() &&
+          State->assume(*CountReached.getAs<DefinedOrUnknownSVal>(), true))
----------------
donat.nagy wrote:
> I think checking the nullness of `getAs()` is more elegant than using a separate `isUndef()` check.
> 
> On a longer term / as a separate improvement, I'd also think about allowing `UndefinedVal` as the argument of the `assert()`-like functions, because the `evalBinOp` -> `assert` combination is very common in checkers and IIRC in most checkers the branch of `UndefinedVal` will produce the same result as `UnknownVal`.
Thanks for the advice!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D158707/new/

https://reviews.llvm.org/D158707