[Bug tree-optimization/128237] [14/15/16/17 Regression] gimple_could_trap_p_1 treats FP-to-integer builtins (lrint, llrint, irint) as non-trapping, vectorized lrint raises FE_INVALID for inactive lanes

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128237

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Target Milestone|---                         |14.5
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2026-10-28
           Priority|P3                          |P2

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  The built-in case should have an explicit list of trapping builtins
that are not catched by the return type.

There's also the general call case that would have to return true, but how IEEE
exception side-effects interact with const/pure attributes is unspecified.