tree 7018026c23fa95b4bbb4baa21d722786ba18ab46
parent a8057ff12956b7dd6dec44f9610c446c828c4af8
author guillem-bartrina-sonarsource <guillem.bartrina@sonarsource.com> 1760430348 +0200
committer GitHub <noreply@github.com> 1760430348 +0200
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJo7gkMCRC1aQ7uu5UhlAAAI9QQACZKCHXtKPvXn2YiU52qUFW4
 cgRHqTGpBW+yv4USfAxaR6f+DtrYveAhI4wWcsvvYwDpANgDrvkn+vE7wKXM2WAf
 wNeu5a+ersCGbZw4wnCFVRVWzMUB4272wrMXuFbJKKL46ivvpVtU7BOAbG4C64t3
 lkxkTAYOnWPi17WBGN7oJqmyL1E0e/na8pBpYjrf8m5Wo7BGmF5nKQNaU8xBuAJ1
 cnmXvD329WMZMhqhMWAUnTsRAQguD6FEU6dNLqBrz7SGX6Wueden2pAwpBkmKCIj
 JyJlR675eO/EMNVnlSxCQY8T4wfntBKSwMmraIXhjALrJvLWtPtZenxfK61URvEd
 bF85xHubCWLddfsYO6c7PHSiM2AfR0z/9XGdBp/OU8vAe0STVFmo+DhLy8wZ3upU
 mjoarXEtCRGw2hEjVvKfBUYqWdwLBlw8pamDjdNZmAtkGQIP1wS3WQKGjfxIp8Ag
 YXwYey6VXKWpfspHNd5NfVuK6wGi1LDVza+8XQp6DDPyyPad2afS8gqME6bxjZeX
 NiftVYUfAF+cKbfx4DELWEfVQPIF0FpXXS3BISMI9HJ+x9y3QrKRnFX2ir6SPWnP
 UGIt8CINfEG25PjvCK1+xTd0TqspPm3p4HwFAt86eN9t/LcHs+jCyMljVqFeeKDW
 YO8DIaQSlWaP1AQ5qVSB
 =fbuO
 -----END PGP SIGNATURE-----
 

[analyzer] `SValBuilder::evalBinOpLN`: try simplifying the RHS first (#161537)

The first thing `SValBuilder::evalBinOpNN` does is simplify both
operators
([here](https://github.com/llvm/llvm-project/blob/main/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp#L430-L437)),
why wouldn't we simplify the RHS in `SValBuilder::evalBinOpLN`?

When the LHS is an `Element`, the right side is simplified incidentally
when calling `evalBinOpNN`
([here](https://github.com/llvm/llvm-project/blob/main/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp#L1169-L1170)).

Without this patch, the “base_complex” test case is **UNKNOWN** instead of **TRUE**.