[PATCH] D68165: [analyzer][MallocChecker][NFC] Split checkPostCall up, deploy CallDescriptionMap

Kristóf Umann via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Wed Oct 2 02:39:24 PDT 2019


Szelethus marked 5 inline comments as done.
Szelethus added inline comments.


================
Comment at: clang/lib/StaticAnalyzer/Checkers/MallocChecker.cpp:378-379
+
+  using CheckFn = void (MallocChecker::*)(CheckerContext &C, const CallExpr *CE,
+                                          ProgramStateRef State) const;
+
----------------
Szelethus wrote:
> NoQ wrote:
> > Whenever i see a (`CE`, `State`) pair, it screams `CallEvent` to me. That said, i'm worried that `State` in these callbacks isn't necessarily equal to `C.getState()` (the latter, by the way, is always equal to the `CallEvent`'s `.getState()` - that's a relief, right?), so if you'll ever be in the mood to check that, that'd be great :)
> It should be always equal to it. I'll change it.
Hmmm, I tried making this take a (`CallEvent`, `CheckerContext`), but it bloated the code for little gain, since every handler function would start with the retrieval of the state and the call expression. That said, I could cascade these down to `FreeMemAux`, `MallocMemAux`, etc, to also take this pair instead, but I'll be honest, I don't see point until something actually breaks.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68165/new/

https://reviews.llvm.org/D68165





More information about the cfe-commits mailing list