[Bug middle-end/109545] New: first_stmt has strange semantics

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109545

            Bug ID: 109545
           Summary: first_stmt has strange semantics
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rguenth at gcc dot gnu.org
  Target Milestone: ---

tree-cfg.cc:first_stmt skips debug stmts but not labels which is quite odd and
can lead to compare-debug issues.

Users are

./tree-cfgcleanup.cc:  stmt = first_stmt (dest);
./tree-cfgcleanup.cc:  label = first_stmt (dest);
./omp-expand.cc:              && first_stmt (store_bb) == store
./tree-vect-slp.cc:     if (gcall *first = safe_dyn_cast <gcall *> (first_stmt
(bb)))
./gimple-loop-versioning.cc:      gimple *stmt = first_stmt (loop->header);
./cp/cp-gimplify.cc:    return first_stmt (p->stmt);
./cp/cp-gimplify.cc:      return first_stmt (BIND_EXPR_BODY (stmt));
./cp/cp-gimplify.cc:      tree ft = first_stmt (then_);
./cp/cp-gimplify.cc:      tree fe = first_stmt (else_);
./tree-cfg.h:extern gimple *first_stmt (basic_block);
./tree-cfg.cc:  then_stmt = first_stmt (then_bb);
./tree-cfg.cc:  else_stmt = first_stmt (else_bb);