[clang] [analyzer] Don't keep a reclaimable node in TrackControlDependencyCondBRVisitor (PR #242711)

Balázs Benics via cfe-commits cfe-commits at lists.llvm.org
Sun Nov 15 03:49:19 PST 2026


steakhal wrote:

Earlier this year in #105512 we discussed that we want to get rid of node reclamation and that it's probably just a final evaluation that holds us back because my #207596 appears to resolved the huge regression that would be caused by the disablement of this reclamation - but I only measured the regressing TUs, so I couldn't tell if the whole regression would go away and it wouldn't introduce new regressions.
Hence I think only a finall evaluation is that holds us back from removing node reclamation and simplifying the code like this.
@NagyDonat @tigbr 

https://github.com/llvm/llvm-project/pull/242711


More information about the cfe-commits mailing list