[PATCH] D58367: [analyzer] NFC: Improve upon the concept of BugReporterVisitor.

[Artem Dergachev via Phabricator via cfe-commits]

NoQ updated this revision to Diff 189948.
NoQ added a comment.

Found a bug! The lambda has to manually make sure that the bug report actually does have something to do with the checker. The attached testcase shows that we definitely need to avoid attaching a MIG checker note to a core.DivideZero bug report. I stuffed a check into the lambda to demonstrate what sort of logic do we need to avoid this problem:

  if (&BR.getBugType() != &BT)
    return "";

But of course i'd much rather automate that because most checkers will need this sort of logic and because it looks really ugly and is very easy to forget. My best idea so far, for the next patch, is to store a pointer to the `Checker` object in both the note and the bug report and automatically compare them in `TagVisitor` before even trying to emit a note. Such check should probably be optional, because otherwise only checkers would be able to use note tags, while i still want the Core to be able to use it as well.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D58367/new/

https://reviews.llvm.org/D58367

Files:
  clang/include/clang/Analysis/ProgramPoint.h
  clang/include/clang/StaticAnalyzer/Core/BugReporter/BugReporterVisitors.h
  clang/include/clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h
  clang/include/clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h
  clang/lib/StaticAnalyzer/Checkers/MIGChecker.cpp
  clang/lib/StaticAnalyzer/Core/BugReporter.cpp
  clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp
  clang/test/Analysis/mig.mm

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D58367.189948.patch
Type: text/x-patch
Size: 11236 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20190308/a4e5c848/attachment-0001.bin>