fix(bindings): CUDA_PYTHON_DISABLE_MAJOR_VERSION_WARNING=0 must not disable it - #2581

Problem

The suppression check tests the raw environment string for truthiness (_version_check.py:35):

    # Allow users to suppress the warning
    if os.environ.get("CUDA_PYTHON_DISABLE_MAJOR_VERSION_WARNING"):
        return

so CUDA_PYTHON_DISABLE_MAJOR_VERSION_WARNING=0 — the obvious way to spell "no, keep warning me" — suppresses the warning exactly as effectively as =1.

Two things make that a trap rather than a convention:

1. The warning's own text reads (Set CUDA_PYTHON_DISABLE_MAJOR_VERSION_WARNING=1 to suppress this warning.), which implies 0 is the off value.
2. The other two boolean knobs in this repository disagree with it, and both parse with int():

# cuda_bindings/cuda/bindings/_internal/runtime_linux.pyx:29 (and _windows)
__usePTDS = bool(int(os.getenv('CUDA_PYTHON_CUDA_PER_THREAD_DEFAULT_STREAM', default=0)))

# cuda_core/cuda/core/__init__.py:44
if int(os.environ.get("CUDA_CORE_DONT_FIX_TAB_COMPLETION", "0")):

A user (or a CI job template, or a Dockerfile) that sets all three to 0 gets the behaviour they asked for from two of them and the opposite from this one — silently losing a compatibility warning whose whole purpose is to explain why their driver is too old for the wheel they installed.

Fix

Parse the value with int() here too, in a small _warning_disabled() helper.

Compatibility: unset and empty still mean "not disabled", every non-zero integer still disables, and a value that is not an integer keeps the old set-means-disabled behaviour so nobody relying on a spelling like =true starts seeing the warning again. 0 is the only input whose meaning changes — which is the point.

Also uses the env-var name constant in the warning text so the message and the lookup cannot drift apart.

Tests

Added to cuda_bindings/tests/test_version_check.py:

• test_disable_flag_parsing — parametrised over 0, 0, empty, blank, 1, 2, true, yes.
• test_disable_flag_unset.
• test_warning_not_suppressed_when_env_var_is_zero — end-to-end through warn_if_cuda_major_version_mismatch, mirroring the existing test_warning_suppressed_by_env_var it sits next to.

The existing test_warning_suppressed_by_env_var (=1) is unchanged and still passes.

What I ran

Environment: macOS, no CUDA driver and no CUDA toolkit, so cuda.bindings is not built here.

• Did not run: cuda_bindings/tests/test_version_check.py itself — it imports cuda.bindings.driver at module level.
• Ran (teeth check): _version_check.py imports only os / threading / warnings at module level and pulls in cuda.bindings.driver inside the function, so I loaded the real module with a stubbed driver (CUDA_VERSION=13000, cuDriverGetVersion -> 12080) and counted emitted warnings for each env value:

                       main    this PR
  unset                  1        1
  ""                     1        1
  "   "                  0        1   <-- fixed
  "0"                    0        1   <-- the defect
  " 0 "                  0        1   <-- fixed
  "1"                    0        0
  "2"                    0        0
  "true"                 0        0   <-- deliberately unchanged
  "yes"                  0        0   <-- deliberately unchanged

• Ran: python -m py_compile, ruff check, ruff format --check on both changed files — clean, no new findings against a main baseline.
• Checked: warn_if_cuda_major_version_mismatch has exactly one non-test caller, Device_ensure_cuda_initialized in cuda_core/cuda/core/_device.pyx:1613-1625, which calls it immediately after a successful cuInit(0) and does not inspect the env var itself.