OpenACC 3.3 gang(dim:) is rejected (C)

// # Finding (GCC 16.2): the OpenACC 3.3 gang(dim:) argument is rejected (C); the reproducer is the catalog program
//   acc34-c-clause-gang-dim
//
// - toolchains: gcc_g90 (GCC 16.2.0, AMD Instinct MI250X gfx90a, Tioga) and gcc_sm80 (GCC 16.2.0,
//   NVIDIA A100 sm_80, wingtip); separate memory only; -O2; OMP_TARGET_OFFLOAD=mandatory
// - build: gcc -fopenacc -foffload=amdgcn-amdhsa -foffload-options=amdgcn-amdhsa=-march=gfx90a -O2 (gcc_g90); -foffload=nvptx-none
//   -foffload-options=nvptx-none=-march=sm_80 -foffload-options=nvptx-none=-latomic -O2 (gcc_sm80)
// - failure: rejected
// - observed: `error: 'dim' undeclared (first use in this function)` on both GPUs.
// - expected: compiles; `RESULT ran_on_host=0 check=pass`
// - spec: the dim argument of gang (OpenACC 3.4 §2.9.2 l.2371; added in 3.3, "Changes from Version 3.2 to 3.3", margin line 800).
// - GCC 16.2 documentation (libgomp.texi "Enabling OpenACC"): "A complete description of all OpenACC directives accepted may be
//   found in the OpenACC Application Programming Interface manual, version 2.6"; _OPENACC is 201711 (2.6). No list of OpenACC gaps
//   is documented.
// - other families: NVHPC 25.11/26.9 compile and run it correctly.
// - gate and confirmer: 4 cells (2 gcc_g90, 2 gcc_sm80), each confirmed in the gate's four environments (a-d;
//   results/gcc/gate-tioga/ and gate-wingtip/ verdicts.jsonl); confirmer verdicts (results/gcc/confirm/batch.gcc_*.jsonl):
//   compiler_bug (R11-compile-rejection) 2, needs_human (R12b-compile-undecided) 2
// - catalog programs (2): acc34-c-clause-gang-dim, acc34-c-clause-gang-dim-binop
// - legality: the catalog programs are legal by construction (offload.engine.gen, c_gen and acc_gen
//   check every program against the catalog rules and the offload scope as it is built; no
//   compiler). GCC is a compiler under test only: never a legality witness, here or for another
//   family's case (catalog/machines.yaml under_test_only).
// Generated by compliance.acc_c_gen (OpenACC 3.4, C11): the C port (offload.engine.f2c) of acc34-clause-gang-dim. Do not edit.
// id: acc34-c-clause-gang-dim
// primary atom: acc.clause.gang.dim
// legal: yes
// observable: yes
// ran_on_host: inside the compute region of interest (acc_on_device)
// evidence: equivalent_default
// needs: none
// std: c11
// oracle: without the loop construct every gang of the parallel region executes every iteration (gang-redundant mode), and the
// atomic update adds the value once per gang
// no oracle (equivalent_default): dimension one is the gang clause's dimension without dim
// basis: OpenACC 3.4 §2.9.2, p.66: "The associated dimension is the value of the dim argument, if it appears, or is dimension one."
// C: the atomic update is written x += expr; the port's x = x + e1 + e2 is also a legal 3.4 §2.12 form (mathematically equivalent
// to x + (e1 + e2)) and is kept as acc34-c-clause-gang-dim-binop
#include <stdio.h>
#include <stdbool.h>
#include <openacc.h>

enum { n = 64 };

int main(void) {
  int a[n];
  int b[n];
  int ref[n];
  int host_flag;
  bool ok;
  for (int i = 1; i <= n; i++) {
    a[i - 1] = i;
    b[i - 1] = 0;
    ref[i - 1] = 2 * i + 1;
  }
  host_flag = 1;
  #pragma acc parallel num_gangs(4) copyin(a) copy(b) copy(host_flag)
  {
    #pragma acc loop gang(dim: 1)
    for (int i = 1; i <= n; i++) {
      if (i == 1) host_flag = (acc_on_device(acc_device_host) ? 1 : 0);
      #pragma acc atomic update
      b[i - 1] += 2 * a[i - 1] + 1;
    }
  }
  bool t1_ = true;
  for (int k2_ = 0; k2_ < n; k2_++)
    t1_ = t1_ && (b[k2_] == ref[k2_]);
  ok = t1_;
  printf("RESULT ran_on_host=%d check=%s\n", host_flag, ok ? "pass" : "fail");
  return 0;
}