[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

This revision was automatically updated to reflect the committed changes. Closed by commit rG9fdcae7c81f5: [analyzer] Do not cache out on some shared implicit AST nodes (authored by xazax.hun). Changed prior to commit: https://reviews.llvm.org/D71371?vs=233468&id=233475#toc Repository: rG LL

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

NoQ accepted this revision. NoQ added a comment. This revision is now accepted and ready to land. Thanks!! Comment at: clang/lib/StaticAnalyzer/Core/ExprEngine.cpp:1318 case Expr::MSDependentExistsStmtClass: llvm_unreachable("Stmt should not be in analyzer evaluatio

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

xazax.hun updated this revision to Diff 233468. xazax.hun added a comment. - Removed leftover code from previous approach. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D71371/new/ https://reviews.llvm.org/D71371 Files: clang/include/clang/Analysis/CFG.h clang/lib/Analysis/CFG.cpp

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

xazax.hun updated this revision to Diff 233467. xazax.hun edited the summary of this revision. xazax.hun added a comment. - Omit nodes from CFG instead of trying to make the states distinct. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D71371/new/ https://reviews.llvm.org/D71371 Files:

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

NoQ added a comment. I doubt we'll ever need these callbacks. I can't imagine any kind of analysis that would need them. Like, these statements have absolutely no effect on the runtime program state. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D7

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

xazax.hun added a comment. Hmm loops. Right. So this solution is insufficient because it can prevent legitimate caching out in loops. If we remove those initializers from the CFG wouldn't we loose the Pre/PostStmt callbacks? Is that acceptable? If not, we might need to create a new program poi

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

NoQ added a comment. One way to criticize the current solution would be to put an `InitListExpr` in a loop and immediately discard it and see how states after different numbers of iterations no longer join together only because they have different `NoCachingOutForConsts` values. Repository:

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

NoQ added a comment. Ahaa. Ahaa. Okay. So the AST is like | |-DeclStmt 0x7f9f3b8932e0 | | `-VarDecl 0x7f9f3b892f50 col:7 used a 'int [5]' cinit | | `-InitListExpr 0x7f9f3b893268 'int [5]' | | |-array_filler: ImplicitValueInitExpr 0x7f9f3b8932d0 <> 'int' | | |-Int

[PATCH] D71371: [analyzer] Do not cache out on some shared implicit AST nodes.

xazax.hun created this revision. xazax.hun added reviewers: NoQ, dcoughlin, Szelethus, baloghadamsoftware, haowei. xazax.hun added a project: clang. Herald added subscribers: Charusso, gamesh411, dkrupp, donat.nagy, mikhail.ramalho, a.sidorin, rnkovacs, szepet. Some AST nodes that stands for imp