Fix bump_build() silently returning an unchanged version for digitless build metadata - #480

Calling Version.parse("1.2.3+alpha").bump_build() returned a version whose build metadata was still "alpha" — an identical version object string-wise, despite bump_build() being documented to raise the build part. The private helper _increment_string() leaves a string without any digits unchanged, so a build like "alpha" (no numeric identifier anywhere) was passed through untouched. bump_build() now detects that case and appends ".0" ("1.2.3+alpha" -> "1.2.3+alpha.0"), exactly mirroring how PR #460 fixed the analogous prerelease case ("3.4.5-rc9" -> "3.4.5-rc9.0"). Numeric-ending builds keep the existing incrementing behavior ("1.2.3+build.9" -> "1.2.3+build.10"). Found while investigating #466 (passing the bump token parameter); the CLI --token option itself is already covered by open PR #467 and is intentionally not duplicated here.

Implementation reason: _increment_string() only increments the last numeric sequence and returns the input unchanged when there is none; for the prerelease part this hole was already fixed in _increment_prerelease by appending ".0" so the bumped version is a new, higher build step. bump_build() had no such guard, so for digitless build metadata it silently returned an equal version, breaking the documented raise contract and making repeated bumps indistinguishable from no-ops. The minimal fix reuses the same ".0"-append strategy at the bump_build() call site, leaving numeric behavior and the token-constructs-new-build behavior untouched.

Validation: a new regression test fails on the unchanged base and passes with this patch; reproduced in an independent clean checkout.

Full-suite results (including pre-existing failures, if any):

{
  "base": {
    "exit": 1,
    "tests": 393,
    "failures": 5,
    "errors": 0,
    "skipped": 0,
    "failed_cases": [
      "tests.test_bump_build_digitless.test_bump_build_digitless_metadata_should_append_0",
      "tests.test_bump_build_digitless.test_bump_build_digitless_metadata_should_return_new_version",
      "tests.test_bump_build_digitless.test_bump_build_digitless_multi_identifier_should_append_0",
      "tests.test_bump_build_digitless.test_bump_build_token_ignored_for_existing_digitless_build",
      "tests.test_bump_build_digitless.test_bump_build_digitless_bump_is_repeatable"
    ]
  },
  "patch": {
    "exit": 0,
    "tests": 393,
    "failures": 0,
    "errors": 0,
    "skipped": 0,
    "failed_cases": []
  }
}

AI assistance: implementation and review used Hermes with self-hosted GLM 5.3. Automated test evidence was checked separately. This does not represent a human review.

Additional checks were reproduced on fresh base and patched checkouts:

These diagnostic counts and samples are also present on the unchanged base; comparison of the complete diagnostic sets found no additions.

{
  "mypy": {
    "total": 1,
    "sample": [
      {
        "file": "src/semver/version.py",
        "code": "import-untyped",
        "message": "Skipping analyzing \"fast_semver_rs_backend\": module is installed, but missing library stubs or py.typed marker"
      }
    ]
  },
  "ruff": {
    "total": 159,
    "sample": [
      {
        "file": "docs/advanced/coerce.py",
        "code": "I001",
        "message": "Import block is un-sorted or un-formatted"
      },
      {
        "file": "docs/advanced/coerce.py",
        "code": "UP035",
        "message": "`typing.Tuple` is deprecated, use `tuple` instead"
      },
      {
        "file": "docs/advanced/coerce.py",
        "code": "UP006",
        "message": "Use `tuple` instead of `Tuple` for type annotation"
      },
      {
        "file": "docs/advanced/coerce.py",
        "code": "UP045",
        "message": "Use `X | None` for type annotations"
      },
      {
        "file": "docs/advanced/coerce.py",
        "code": "RUF100",
        "message": "Unused `noqa` directive (non-enabled: `E203`)"
      }
    ]
  },
  "docformatter": {
    "total": 24,
    "sample": [
      {
        "file": "before/src/semver/_deprecated.py",
        "code": "docformat",
        "message": "-    :param replace: the function to replace (use the full qualified"
      },
      {
        "file": "before/src/semver/_deprecated.py",
        "code": "docformat",
        "message": "-        name like ``semver.version.Version.bump_major``."
      },
      {
        "file": "before/src/semver/_deprecated.py",
        "code": "docformat",
        "message": "+    :param replace: the function to replace (use the full qualified name like"
      },
      {
        "file": "before/src/semver/_deprecated.py",
        "code": "docformat",
        "message": "+        ``semver.version.Version.bump_major``."
      },
      {
        "file": "before/src/semver/_deprecated.py",
        "code": "docformat",
        "message": "-    :param category: allow you to specify the deprecation warning class"
      }
    ]
  }
}

Changelog: the reviewed bugfix entry is attached using this PR number after draft creation, following changelog.d/README.rst. The PR is marked ready only after the fragment passes Towncrier.