steakhal added a comment. I found this report at `vim/src/term.c` (https://github.com/vim/vim.git at `v8.2.1920`): F23102904: image.png <https://reviews.llvm.org/F23102904>
I'm not sure what the exact type of `BC` global variable is but I think it's declared as either `extern char *BC;` or just `char *BC;`. `empty_option` is declared as either `extern unsigned char *empty_option;` or `unsigned char *empty_option = (unsigned char *)"";` Could you please have a look at this FP @zukatsinadze? ================ Comment at: clang/lib/StaticAnalyzer/Checkers/CMakeLists.txt:68-70 MIGChecker.cpp + cert/ModelConstQualifiedReturnChecker.cpp MoveChecker.cpp ---------------- zukatsinadze wrote: > steakhal wrote: > > Put this in the right alphabetical place. > I think it is in the right alphabetical place not considering cert/ (other > certs are in similar order), but maybe I should remove the cert directory, > what do you think? > It was created by me a few years ago, but I don't see a point now anymore. Leave it here. ================ Comment at: clang/lib/StaticAnalyzer/Checkers/StoreToImmutableChecker.cpp:52 + ImmutableMemoryBind, "modifying immutable memory", ErrorNode); + Report->markInteresting(R); + C.emitReport(std::move(Report)); ---------------- zukatsinadze wrote: > NoQ wrote: > > I also recommend `trackExpressionValue` to make sure we have all > > reassignments highlighted as the value gets bounced between pointer > > variables. The user will need proof that the pointer actually points to > > const memory. > What expression should I use for tracking? I tried to simply cast `S` to > `Expr`, but it didn't really work. > Then I came up with something ugly: cast `S` to `BinaryOperator` -> `getLHS` > -> `getSubExpr` -> `ignoreImpCasts`. Transitive interestingness might be related to this: D125362 CHANGES SINCE LAST ACTION https://reviews.llvm.org/D124244/new/ https://reviews.llvm.org/D124244 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
