crapkit 0.7.4__tar.gz → 0.7.6__tar.gz

{crapkit-0.7.4/src/crapkit.egg-info → crapkit-0.7.6}/PKG-INFO

@@ -1,11 +1,11 @@
 Metadata-Version: 2.4
 Name: crapkit
-Version: 0.7.4
+Version: 0.7.6
 Summary: Scores every function on complexity times uncovered risk, ranks the worst, and blocks commits that add more.
 Author: Jean-Francois Gagne
 License: MIT
 Project-URL: Homepage, https://github.com/JeanFrancoisGagne/crapkit
-Project-URL: Documentation, https://jeanfrancoisgagne.github.io/crapkit/handbook.html
+Project-URL: Documentation, https://www.jfgagne.com/crapkit/handbook.html
 Project-URL: Changelog, https://github.com/JeanFrancoisGagne/crapkit/blob/main/CHANGELOG.md
 Project-URL: Issues, https://github.com/JeanFrancoisGagne/crapkit/issues
 Project-URL: Source, https://github.com/JeanFrancoisGagne/crapkit
@@ -111,6 +111,9 @@ crapkit ratchet seed
 git add crapkit.toml crapkit-ratchet.tsv .gitignore
 ```
 
+Not a Python repo? `uvx crapkit init` runs the same commands and adds nothing to your
+manifest: see [A repo that is not Python](#a-repo-that-is-not-python).
+
 `init` detects pytest, Vitest and Jest from the repository's own files. Review the
 generated config before running its commands. When detection leaves a commented
 lane, fill it in using the [lane recipes](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md).
@@ -182,6 +185,28 @@ pip install git+https://github.com/JeanFrancoisGagne/crapkit.git
 pip install .
 ```
 
+### A repo that is not Python
+
+crapkit is a command-line tool, never a dependency of the code it scores. A TypeScript,
+Go or Rust repo adds nothing to its own manifest. With [uv](https://docs.astral.sh/uv/)
+on the machine, `uvx` fetches crapkit into a cache of its own and runs it:
+
+```
+$ uvx crapkit init
+wrote crapkit.toml with 1 scope(s): src
+detected 1 lane(s) from this repo's own files: js - next: run `crapkit coverage`
+added to .gitignore: .crapkit/
+
+$ uvx crapkit coverage
+$ uvx crapkit worklist
+```
+
+The lane still runs your own test runner, so Vitest or Jest and its coverage package come
+from the repo's `node_modules` as they do today. `uv tool install crapkit` or
+`pipx install crapkit` puts a `crapkit` command on PATH once, which is what the
+[commit gate](#the-gate) and the Claude Code plugin call. uv brings its own Python when
+the machine has none.
+
 Requires Python 3.11 or newer and Git on PATH. The CLI has one runtime dependency,
 `lizard>=1.24.0`; a package mirror needs both distributions. Install into the environment
 you intend to use, then check `crapkit --version`. The `pip install -e ".[dev]"` under
@@ -199,7 +224,7 @@ files. Review those commands before running Crapkit in a repository you do not t
 
 ```
 $ crapkit --version
-crapkit 0.7.4
+crapkit 0.7.6
 ```
 
 `python -m crapkit` works identically to the console script and is what to use from a
@@ -455,7 +480,7 @@ crapkit ships a `.pre-commit-hooks.yaml` declaring `id: crapkit-gate`. In your
 repos:
   - repo: https://github.com/JeanFrancoisGagne/crapkit
     # crapkit's release step rewrites this line to the tag it just cut
-    rev: v0.7.4
+    rev: v0.7.6
     hooks:
       - id: crapkit-gate
 ```
@@ -563,7 +588,7 @@ to a workflow, and every input has a default:
       - uses: actions/checkout@v4
         with:
           fetch-depth: 0
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
 ```
 
 The whole job those four lines sit in:
@@ -583,7 +608,7 @@ jobs:
         with:
           python-version: "3.12"       # the interpreter the install below lands in
       - run: pip install -e ".[dev]"   # whatever your lanes need to run
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
         with:
           gate: "false"
 ```
@@ -654,7 +679,7 @@ closes it. A verify that passed is one line: `**verify passed.** Run 2 against b
 The rows are the ranked worklist for the files the pull request changed, worst first,
 `top` of them, with the rows a finding names listed first. `risk` is ccn times churn
 weight, the number `crapkit worklist` ranks on, and `remedy` is the run's own verdict for
-that function: `decompose`, `add-tests` or `ok`. `(accepted debt)` marks a function the
+that function: `decompose`, `split-lines`, `add-tests` or `ok`. `(accepted debt)` marks a function the
 committed ratchet carries a mark for, so an untouched `legacy_router` does not read like
 the pull request's own new function. A pull request that touches no ranked function gets
 the heading and no table.
@@ -830,6 +855,7 @@ The coverage summary counts all four as `measured` / `untested` / `no_lane` / `c
 | Remedy | Condition | Action |
 |---|---|---|
 | `decompose` | `ccn > ceiling` | Split it. No amount of coverage clears this. |
+| `split-lines` | `ccn <= ceiling`, `crap > ceiling`, and another function shares its source lines | Put each definition on its own lines, then measure again. Coverage cannot tell functions on one line apart, so the score stays at uncovered whatever the tests do. |
 | `add-tests` | `ccn <= ceiling` and `crap > ceiling` | Cover the branches. |
 | `ok` | `crap <= ceiling` | Nothing. |
 
@@ -1268,12 +1294,12 @@ export function classify(row: Row): string {
 ```
 $ crapkit rescore src/grade.ts --gate
 rescore vs run 1 @ 8bfbe613fcd (coverage STALE, complexity fresh)
-   ccn   cov     crap  remedy     function
-     5    0%     30.0  add-tests  src/grade.ts:22  band ( score )
-     5    0%     30.0  add-tests  src/grade.ts:38  demote ( letter , row Row )
-     4    0%     20.0  add-tests  src/grade.ts:8  penalty ( attempts , late )
-     4   45%      6.7  add-tests  src/grade.ts:48  classify ( row Row )
-     4   75%      4.2  ok         src/grade.ts:59  average ( scores Array )
+   ccn   cov     crap  remedy      function
+     5    0%     30.0  add-tests   src/grade.ts:22  band ( score )
+     5    0%     30.0  add-tests   src/grade.ts:38  demote ( letter , row Row )
+     4    0%     20.0  add-tests   src/grade.ts:8  penalty ( attempts , late )
+     4   45%      6.7  add-tests   src/grade.ts:48  classify ( row Row )
+     4   75%      4.2  ok          src/grade.ts:59  average ( scores Array )
 ```
 
 Exit 0: every piece is at or under 6. The `crap` column is loud because its coverage half
@@ -1340,7 +1366,7 @@ with no debt.
 
 | Page | Covers |
 |---|---|
-| [The handbook](https://jeanfrancoisgagne.github.io/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://jeanfrancoisgagne.github.io/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
+| [The handbook](https://www.jfgagne.com/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://www.jfgagne.com/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
 | [docs/adoption.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/adoption.md) | The judgment layer over the quickstarts: scope granularity, exclude vs lane, scoped_tests wiring, the first-verify taint hazard. |
 | [docs/configuration.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/configuration.md) | Every `crapkit.toml` key: type, default, and what it does. |
 | [docs/lanes.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md) | The lane model, vitest and jest and pytest recipes, artifact reuse, flake retest, containers. |
@@ -1371,6 +1397,12 @@ line arms the complexity gate. See
 for development and [the verified implementation report](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/architecture/2026-09-07-implementation/REPORT.md)
 for complete Windows source and Linux wheel results, focused benchmarks and their limits.
 
+## Maintainer and project background
+
+crapkit is created and maintained by [Jean-François Gagné](https://www.jfgagne.com/).
+Read the [project background](https://www.jfgagne.com/projects/crapkit/) for the
+problem it addresses and how it fits into his work on software and AI.
+
 ## License
 
 MIT. See [LICENSE](https://github.com/JeanFrancoisGagne/crapkit/blob/main/LICENSE).

{crapkit-0.7.4 → crapkit-0.7.6}/README.md

@@ -70,6 +70,9 @@ crapkit ratchet seed
 git add crapkit.toml crapkit-ratchet.tsv .gitignore
 ```
 
+Not a Python repo? `uvx crapkit init` runs the same commands and adds nothing to your
+manifest: see [A repo that is not Python](#a-repo-that-is-not-python).
+
 `init` detects pytest, Vitest and Jest from the repository's own files. Review the
 generated config before running its commands. When detection leaves a commented
 lane, fill it in using the [lane recipes](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md).
@@ -141,6 +144,28 @@ pip install git+https://github.com/JeanFrancoisGagne/crapkit.git
 pip install .
 ```
 
+### A repo that is not Python
+
+crapkit is a command-line tool, never a dependency of the code it scores. A TypeScript,
+Go or Rust repo adds nothing to its own manifest. With [uv](https://docs.astral.sh/uv/)
+on the machine, `uvx` fetches crapkit into a cache of its own and runs it:
+
+```
+$ uvx crapkit init
+wrote crapkit.toml with 1 scope(s): src
+detected 1 lane(s) from this repo's own files: js - next: run `crapkit coverage`
+added to .gitignore: .crapkit/
+
+$ uvx crapkit coverage
+$ uvx crapkit worklist
+```
+
+The lane still runs your own test runner, so Vitest or Jest and its coverage package come
+from the repo's `node_modules` as they do today. `uv tool install crapkit` or
+`pipx install crapkit` puts a `crapkit` command on PATH once, which is what the
+[commit gate](#the-gate) and the Claude Code plugin call. uv brings its own Python when
+the machine has none.
+
 Requires Python 3.11 or newer and Git on PATH. The CLI has one runtime dependency,
 `lizard>=1.24.0`; a package mirror needs both distributions. Install into the environment
 you intend to use, then check `crapkit --version`. The `pip install -e ".[dev]"` under
@@ -158,7 +183,7 @@ files. Review those commands before running Crapkit in a repository you do not t
 
 ```
 $ crapkit --version
-crapkit 0.7.4
+crapkit 0.7.6
 ```
 
 `python -m crapkit` works identically to the console script and is what to use from a
@@ -414,7 +439,7 @@ crapkit ships a `.pre-commit-hooks.yaml` declaring `id: crapkit-gate`. In your
 repos:
   - repo: https://github.com/JeanFrancoisGagne/crapkit
     # crapkit's release step rewrites this line to the tag it just cut
-    rev: v0.7.4
+    rev: v0.7.6
     hooks:
       - id: crapkit-gate
 ```
@@ -522,7 +547,7 @@ to a workflow, and every input has a default:
       - uses: actions/checkout@v4
         with:
           fetch-depth: 0
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
 ```
 
 The whole job those four lines sit in:
@@ -542,7 +567,7 @@ jobs:
         with:
           python-version: "3.12"       # the interpreter the install below lands in
       - run: pip install -e ".[dev]"   # whatever your lanes need to run
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
         with:
           gate: "false"
 ```
@@ -613,7 +638,7 @@ closes it. A verify that passed is one line: `**verify passed.** Run 2 against b
 The rows are the ranked worklist for the files the pull request changed, worst first,
 `top` of them, with the rows a finding names listed first. `risk` is ccn times churn
 weight, the number `crapkit worklist` ranks on, and `remedy` is the run's own verdict for
-that function: `decompose`, `add-tests` or `ok`. `(accepted debt)` marks a function the
+that function: `decompose`, `split-lines`, `add-tests` or `ok`. `(accepted debt)` marks a function the
 committed ratchet carries a mark for, so an untouched `legacy_router` does not read like
 the pull request's own new function. A pull request that touches no ranked function gets
 the heading and no table.
@@ -789,6 +814,7 @@ The coverage summary counts all four as `measured` / `untested` / `no_lane` / `c
 | Remedy | Condition | Action |
 |---|---|---|
 | `decompose` | `ccn > ceiling` | Split it. No amount of coverage clears this. |
+| `split-lines` | `ccn <= ceiling`, `crap > ceiling`, and another function shares its source lines | Put each definition on its own lines, then measure again. Coverage cannot tell functions on one line apart, so the score stays at uncovered whatever the tests do. |
 | `add-tests` | `ccn <= ceiling` and `crap > ceiling` | Cover the branches. |
 | `ok` | `crap <= ceiling` | Nothing. |
 
@@ -1227,12 +1253,12 @@ export function classify(row: Row): string {
 ```
 $ crapkit rescore src/grade.ts --gate
 rescore vs run 1 @ 8bfbe613fcd (coverage STALE, complexity fresh)
-   ccn   cov     crap  remedy     function
-     5    0%     30.0  add-tests  src/grade.ts:22  band ( score )
-     5    0%     30.0  add-tests  src/grade.ts:38  demote ( letter , row Row )
-     4    0%     20.0  add-tests  src/grade.ts:8  penalty ( attempts , late )
-     4   45%      6.7  add-tests  src/grade.ts:48  classify ( row Row )
-     4   75%      4.2  ok         src/grade.ts:59  average ( scores Array )
+   ccn   cov     crap  remedy      function
+     5    0%     30.0  add-tests   src/grade.ts:22  band ( score )
+     5    0%     30.0  add-tests   src/grade.ts:38  demote ( letter , row Row )
+     4    0%     20.0  add-tests   src/grade.ts:8  penalty ( attempts , late )
+     4   45%      6.7  add-tests   src/grade.ts:48  classify ( row Row )
+     4   75%      4.2  ok          src/grade.ts:59  average ( scores Array )
 ```
 
 Exit 0: every piece is at or under 6. The `crap` column is loud because its coverage half
@@ -1299,7 +1325,7 @@ with no debt.
 
 | Page | Covers |
 |---|---|
-| [The handbook](https://jeanfrancoisgagne.github.io/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://jeanfrancoisgagne.github.io/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
+| [The handbook](https://www.jfgagne.com/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://www.jfgagne.com/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
 | [docs/adoption.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/adoption.md) | The judgment layer over the quickstarts: scope granularity, exclude vs lane, scoped_tests wiring, the first-verify taint hazard. |
 | [docs/configuration.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/configuration.md) | Every `crapkit.toml` key: type, default, and what it does. |
 | [docs/lanes.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md) | The lane model, vitest and jest and pytest recipes, artifact reuse, flake retest, containers. |
@@ -1330,6 +1356,12 @@ line arms the complexity gate. See
 for development and [the verified implementation report](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/architecture/2026-09-07-implementation/REPORT.md)
 for complete Windows source and Linux wheel results, focused benchmarks and their limits.
 
+## Maintainer and project background
+
+crapkit is created and maintained by [Jean-François Gagné](https://www.jfgagne.com/).
+Read the [project background](https://www.jfgagne.com/projects/crapkit/) for the
+problem it addresses and how it fits into his work on software and AI.
+
 ## License
 
 MIT. See [LICENSE](https://github.com/JeanFrancoisGagne/crapkit/blob/main/LICENSE).

{crapkit-0.7.4 → crapkit-0.7.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "crapkit"
-version = "0.7.4"
+version = "0.7.6"
 description = "Scores every function on complexity times uncovered risk, ranks the worst, and blocks commits that add more."
 readme = { file = "README.md", content-type = "text/markdown" }
 license = { text = "MIT" }
@@ -49,7 +49,7 @@ dependencies = [
 
 [project.urls]
 Homepage = "https://github.com/JeanFrancoisGagne/crapkit"
-Documentation = "https://jeanfrancoisgagne.github.io/crapkit/handbook.html"
+Documentation = "https://www.jfgagne.com/crapkit/handbook.html"
 Changelog = "https://github.com/JeanFrancoisGagne/crapkit/blob/main/CHANGELOG.md"
 Issues = "https://github.com/JeanFrancoisGagne/crapkit/issues"
 Source = "https://github.com/JeanFrancoisGagne/crapkit"

{crapkit-0.7.4 → crapkit-0.7.6}/src/crapkit/__init__.py

@@ -1,2 +1,2 @@
 """crapkit: deterministic CRAP-score framework."""
-__version__ = "0.7.4"
+__version__ = "0.7.6"

{crapkit-0.7.4 → crapkit-0.7.6}/src/crapkit/cli/scoring.py

@@ -281,7 +281,7 @@ def _scored_run(root: Path, cfg, lanes, *, reuse_artifacts: bool, reuse_unchange
     `git` is the caller's GitFacts when it already has one — verify asks for the
     dirty set before this runs, and that answer is the one the lanes must see too.
     """
-    from ..score import score_rows
+    from ..score import SharedSpanFold, score_rows
 
     git = git or GitFacts(root)
     commit, rows, corpus, cache_hits, tool_versions = _build_inventory(root, cfg, git)
@@ -294,14 +294,55 @@ def _scored_run(root: Path, cfg, lanes, *, reuse_artifacts: bool, reuse_unchange
     # Only scopes a SUCCESSFUL lane covers count as measured; a failed lane's
     # scopes fall back to no-lane flags rather than reading as untested code.
     lane_scopes = {s for lane in succeeded for s in lane.scopes}
+    shared_spans = SharedSpanFold()
     scored = score_rows(rows, coverage_by_path, lane_scopes=lane_scopes, target=cfg.target,
                         scope_targets=cfg.scope_targets,
-                        cc_only_scopes=cfg.coverage_optional_scopes)
+                        cc_only_scopes=cfg.coverage_optional_scopes,
+                        shared_spans=shared_spans)
+    _note_shared_spans(shared_spans, cfg)
     test_failures = {f for prov in provenance.values() for f in prov.get("failures", ())}
     return _ScoredRun(commit, scored, provenance, lane_errors, test_failures, tool_versions,
                       corpus, cache_hits, dead_lines)
 
 
+_SPANS_NAMED = 3
+
+
+def _over_target_at_zero(members, cfg):
+    """The worst function on each shared span that its ceiling can fail.
+
+    Every function on a shared span scores as uncovered, and a function of
+    complexity 2 scores 6 at zero coverage, so no ceiling of 6 can fail it.
+    The count carries those; the named lines are the ones worth splitting.
+    """
+    from ..score import crap
+
+    for span in members:
+        worst = max(span, key=lambda row: row.ccn)
+        if crap(worst.ccn, 0.0) > cfg.ceiling_of(worst.scope):
+            yield worst
+
+
+def _note_shared_spans(fold, cfg) -> None:
+    """Name the source line spans more than one function declares.
+
+    Loud but not fatal, the shape `_note_unanalyzable` settled for unreadable
+    files: the join cannot tell whose coverage is whose on such a span, and
+    ending the run over it left the consumer repo with no coverage run at all.
+    """
+    if not fold.sites:
+        return
+    over = sorted(_over_target_at_zero(fold.sites, cfg), key=lambda row: (-row.ccn, row.path))
+    tail = f"; {len(over)} of them hold a function over its target:" if over else ""
+    print(f"crapkit: {len(fold.sites)} source line span(s) hold more than one function; "
+          "coverage cannot say whose is whose, so those functions score as uncovered, and "
+          f"splitting the definitions onto separate lines measures them{tail}", file=sys.stderr)
+    for row in over[:_SPANS_NAMED]:
+        print(f"crapkit:   {row.path}:{row.start} (complexity {row.ccn})", file=sys.stderr)
+    if len(over) > _SPANS_NAMED:
+        print(f"crapkit:   ... and {len(over) - _SPANS_NAMED} more", file=sys.stderr)
+
+
 def _run_kind(lanes, cfg, failures) -> str:
     """A --lane subset or a run with failed lanes must never serve as the
     verify baseline: its lane set differs from what verify runs, so every
@@ -776,6 +817,6 @@ def cmd_rescore(args: argparse.Namespace) -> int:
 def _print_rescore_table(overlay, latest: dict) -> None:
     """The refactor loop's view: fresh ccn, worst first, stale cov labeled."""
     print(f"rescore vs run {latest['id']} @ {latest['commit'][:11]} (coverage STALE, complexity fresh)")
-    print(f"  {'ccn':>4} {'cov':>5} {'crap':>8}  {'remedy':10} function")
+    print(f"  {'ccn':>4} {'cov':>5} {'crap':>8}  {'remedy':11} function")
     for r in sorted(overlay, key=lambda x: (-x.ccn, x.path, x.start)):
-        print(f"  {r.ccn:>4} {r.cov:>5.0%} {r.crap:>8.1f}  {r.remedy:10} {r.path}:{r.start}  {r.long_name}")
+        print(f"  {r.ccn:>4} {r.cov:>5.0%} {r.crap:>8.1f}  {r.remedy:11} {r.path}:{r.start}  {r.long_name}")

{crapkit-0.7.4 → crapkit-0.7.6}/src/crapkit/mcp_server.py

@@ -30,6 +30,15 @@ _NAME_DESCRIPTION = ("the bare identifier (classify, or route for a Rust "
                      "printed (classify( score , late )); both resolve, exact "
                      "match first")
 
+# The scorer's remedy vocabulary, said once: a structured result carrying a value
+# its outputSchema does not list is rejected whole by a validating client.
+_REMEDIES = ("decompose", "split-lines", "add-tests", "ok")
+_REMEDY_DESCRIPTION = ("decompose (ccn over ceiling), split-lines (another function shares its "
+                       "source lines, so coverage cannot tell them apart and no test lowers the "
+                       "score until the definitions sit on separate lines), add-tests (coverage "
+                       "short) or ok (nothing left to do)")
+_REMEDY = {"type": "string", "description": _REMEDY_DESCRIPTION, "enum": _REMEDIES}
+
 # The partition a large repo needs before `top` means anything: one --scope
 # per element, exact names as declared in crapkit.toml.
 _SCOPE = {
@@ -63,10 +72,7 @@ _PACKET_PROPERTIES = {'scope': {'type': 'string', 'description': 'the declared s
                          'measure this span',
           'enum': ('measured', 'untested', 'no-lane', 'cc-only')},
  'crap': {'type': 'number', 'description': 'the score: ccn^2 x (1 - cov)^3 + ccn'},
- 'remedy': {'type': 'string',
-            'description': 'decompose (ccn over ceiling), add-tests (coverage short) or ok '
-                           '(nothing left to do)',
-            'enum': ('decompose', 'add-tests', 'ok')},
+ 'remedy': _REMEDY,
  'target': {'type': 'integer', 'description': "this scope's effective ccn ceiling"},
  'commits': {'type': 'integer', 'description': 'commits touching the file in the churn window'},
  'authors': {'type': 'integer', 'description': 'distinct authors of those commits'},
@@ -112,9 +118,9 @@ _WORKLIST_ITEM = {'type': 'object',
                                         'inventory-only run',
                          'enum': ('measured', 'untested', 'no-lane', 'cc-only', None)},
                 'remedy': {'type': 'string',
-                           'description': 'decompose, add-tests or ok; only decompose and '
-                                          'add-tests rows with a lane reach get_next_item',
-                           'enum': ('decompose', 'add-tests', 'ok')},
+                           'description': 'decompose, split-lines, add-tests or ok; every row '
+                                          'but ok reaches get_next_item when a lane measures it',
+                           'enum': _REMEDIES},
                 'crap': {'type': ('number', 'null'),
                          'description': 'the score from the ranked run; null on an inventory-only '
                                         'run'},
@@ -469,9 +475,9 @@ TOOLS: tuple[dict, ...] = (
                 "value get_next_item prints")},
             "remedy": {
                 "type": "string",
-                "description": ("decompose, add-tests or ok: the branch the session takes; "
-                "scored.remedy carries the same value"),
-                "enum": ("decompose", "add-tests", "ok")},
+                "description": ("decompose, split-lines, add-tests or ok: the branch the "
+                "session takes; scored.remedy carries the same value"),
+                "enum": _REMEDIES},
             "target": {
                 "type": "integer",
                 "description": ("this scope's effective ccn ceiling, the same value as "
@@ -527,11 +533,7 @@ TOOLS: tuple[dict, ...] = (
                     "crap": {
                         "type": "number",
                         "description": "the score: ccn^2 x (1 - cov)^3 + ccn"},
-                    "remedy": {
-                        "type": "string",
-                        "description": ("decompose (ccn over ceiling), add-tests (coverage short) or "
-                        "ok (nothing left to do)"),
-                        "enum": ("decompose", "add-tests", "ok")}}},
+                    "remedy": _REMEDY}},
             "source": {
                 "type": "string",
                 "description": ("the function's own text, start to end inclusive, newlines intact: "
@@ -590,11 +592,7 @@ TOOLS: tuple[dict, ...] = (
                         "crap": {
                             "type": "number",
                             "description": "score"},
-                        "remedy": {
-                            "type": "string",
-                            "description": ("decompose (ccn over ceiling), add-tests (coverage short) "
-                            "or ok (nothing left to do)"),
-                            "enum": ("decompose", "add-tests", "ok")}}}},
+                        "remedy": _REMEDY}}},
             "file_totals": {
                 "type": "object",
                 "description": "the file rolled up",
@@ -1342,11 +1340,7 @@ TOOLS: tuple[dict, ...] = (
                         "crap": {
                             "type": "number",
                             "description": "score from fresh ccn and baseline cov"},
-                        "remedy": {
-                            "type": "string",
-                            "description": ("decompose (ccn over ceiling), add-tests (coverage short) "
-                            "or ok (nothing left to do)"),
-                            "enum": ("decompose", "add-tests", "ok")},
+                        "remedy": _REMEDY,
                         "stale_coverage": {
                             "type": "boolean",
                             "description": ("always true: complexity is the working tree's, coverage "
@@ -1393,11 +1387,7 @@ TOOLS: tuple[dict, ...] = (
                                 "crap": {
                                     "type": "number",
                                     "description": "score"},
-                                "remedy": {
-                                    "type": "string",
-                                    "description": ("decompose (ccn over ceiling), add-tests "
-                                    "(coverage short) or ok (nothing left to do)"),
-                                    "enum": ("decompose", "add-tests", "ok")},
+                                "remedy": _REMEDY,
                                 "key_name": {
                                     "type": "string",
                                     "description": "the ratchet key form of the name"},

{crapkit-0.7.4 → crapkit-0.7.6}/src/crapkit/score.py

@@ -128,13 +128,20 @@ def _best_match(row: InventoryRow, candidates: list[FnCoverage]) -> FnCoverage |
     return best
 
 
-def _remedy(ccn: int, score: float, ceiling: int) -> str:
+def _remedy(ccn: int, score: float, ceiling: int, shared_span: bool = False) -> str:
+    """The first thing that can lower this score. A function sharing its source
+    line span with another scores as uncovered whatever its tests do, so
+    add-tests there is advice nobody can follow: splitting the definitions is,
+    and the run after the split says whether tests are still owed."""
     if ccn > ceiling:
         return "decompose"
-    return "ok" if score <= ceiling else "add-tests"
+    if score <= ceiling:
+        return "ok"
+    return "split-lines" if shared_span else "add-tests"
 
 
-def _finish(row, cov: float, flag: str, *, target: int, scope_targets) -> ScoredRow:
+def _finish(row, cov: float, flag: str, *, target: int, scope_targets,
+            shared_span: bool = False) -> ScoredRow:
     # cc-only is the pre-commit hook's rule: crap IS ccn, so _remedy can only
     # answer ok or decompose. Feeding it cov=0 through the formula would say
     # add-tests about code no test can reach.
@@ -146,7 +153,7 @@ def _finish(row, cov: float, flag: str, *, target: int, scope_targets) -> Scored
     # built a throwaway dict per row and looked every field up by name.
     return ScoredRow(row[0], row[1], row[2], row[3], row[4], row[5], row[6], row[7],
                      row[8], row[9], row[10],
-                     cov, flag, score, _remedy(row[7], score, ceiling), row[11], row[12])
+                     cov, flag, score, _remedy(row[7], score, ceiling, shared_span), row[11], row[12])
 
 
 def _nearest_overlay(row, candidates) -> list:
@@ -277,35 +284,87 @@ def overlay_stale_coverage(
         if r.flag == "measured":
             by_key.setdefault((r.path, r.long_name), []).append(r)
 
+    shared = _shared_source_spans(rows, lane_scopes, cc_only_scopes)
     scored = []
     for row in rows:
-        cov, flag = (_cov_without_join(row, lane_scopes, cc_only_scopes)
-                     or _named_overlay_cov(row, by_key, positions))
-        scored.append(_finish(row, cov, flag, target=target, scope_targets=scope_targets))
+        verdict = _cov_without_join(row, lane_scopes, cc_only_scopes)
+        cov, flag = verdict or _named_overlay_cov(row, by_key, positions)
+        scored.append(_finish(row, cov, flag, target=target, scope_targets=scope_targets,
+                              shared_span=_on_shared_span(row, verdict, shared)))
     return scored
 
 
+class SharedSpanFold:
+    """The source line spans more than one function declares in one run.
+
+    The join cannot tell whose coverage is whose there, so each such function
+    scores as uncovered and the run names the spans instead of ending. Filled
+    while scoring; the caller reports it once.
+    """
+
+    def __init__(self) -> None:
+        self.sites: list[list[InventoryRow]] = []
+
+    def add(self, members: list[InventoryRow]) -> None:
+        self.sites.append(members)
+
+
+def _identity(row) -> tuple:
+    """What separates two functions on one span: the name, and the occurrence
+    that tells sibling callbacks on a line apart."""
+    return row.long_name, row.occurrence
+
+
+def _join_member(members: list, row) -> None:
+    """A third function on the span joins its members; another copy of one that
+    is already there (the same function in a second scope) does not."""
+    if all(_identity(member) != _identity(row) for member in members):
+        members.append(row)
+
+
 def _shared_source_spans(rows, lane_scopes: set, cc_only_scopes) -> dict:
-    seen, collisions = {}, {}
+    """span -> the distinct functions declaring it, for spans more than one does."""
+    first, collisions = {}, {}
     for row in rows:
         if _cov_without_join(row, lane_scopes, cc_only_scopes) is not None:
             continue
         span = row.path, row.start, row.end
-        identity = row.long_name, row.occurrence
-        if seen.setdefault(span, identity) != identity:
-            collisions[span] = row
+        seen = first.setdefault(span, row)
+        if _identity(seen) != _identity(row):
+            _join_member(collisions.setdefault(span, [seen]), row)
     return collisions
 
 
-def _check_coverage_spans(rows, coverage_by_path: dict, start_index: dict,
-                         lane_scopes: set, cc_only_scopes) -> None:
-    for row in _shared_source_spans(rows, lane_scopes, cc_only_scopes).values():
-        _, flag = _span_join_cov(row, coverage_by_path, start_index)
-        if flag == "measured":
-            raise ToolError(
-                f"coverage for {row.path}:{row.start} cannot distinguish functions "
-                "with the same line span; split their definitions onto separate "
-                "lines and regenerate coverage")
+def _ambiguous_spans(shared: dict, coverage_by_path: dict, start_index: dict) -> dict:
+    """The shared spans a measurement would speak about, so the join would hand
+    every function on the span one function's number.
+
+    Through 0.7.4 this raised and ended the run. One consumer repo holds 591
+    shared spans, 459 of them measured, so a run died on the first one it met
+    and the repo never finished a coverage run at all. Refusing the ambiguous
+    number is the specified behaviour; ending the run over it was not, the
+    shape _note_unanalyzable settled for unreadable files.
+    """
+    ambiguous = {}
+    for span, members in shared.items():
+        if _span_join_cov(members[0], coverage_by_path, start_index)[1] == "measured":
+            ambiguous[span] = members
+    return ambiguous
+
+
+def _joined_cov(row, ambiguous: dict, coverage_by_path: dict,
+                start_index: dict) -> tuple[float, str]:
+    """Uncovered on an ambiguous span, never the neighbour's number: the honest
+    floor for a function whose measurement cannot be told from another's."""
+    if (row.path, row.start, row.end) in ambiguous:
+        return 0.0, "untested"
+    return _span_join_cov(row, coverage_by_path, start_index)
+
+
+def _on_shared_span(row, verdict, shared: dict) -> bool:
+    """A row some lane measures, on a span it shares. Measured yet or not: tests
+    would only make the span measured, and then it scores as uncovered."""
+    return verdict is None and (row.path, row.start, row.end) in shared
 
 
 def score_rows(
@@ -316,12 +375,18 @@ def score_rows(
     target: int = 6,
     scope_targets: dict[str, int] | None = None,
     cc_only_scopes: frozenset[str] = frozenset(),
+    shared_spans: SharedSpanFold | None = None,
 ) -> list[ScoredRow]:
     start_index = _start_index(coverage_by_path)
-    _check_coverage_spans(rows, coverage_by_path, start_index, lane_scopes, cc_only_scopes)
+    shared = _shared_source_spans(rows, lane_scopes, cc_only_scopes)
+    ambiguous = _ambiguous_spans(shared, coverage_by_path, start_index)
+    for members in ambiguous.values():
+        if shared_spans is not None:
+            shared_spans.add(members)
     scored = []
     for r in rows:
-        cov, flag = (_cov_without_join(r, lane_scopes, cc_only_scopes)
-                     or _span_join_cov(r, coverage_by_path, start_index))
-        scored.append(_finish(r, cov, flag, target=target, scope_targets=scope_targets))
+        verdict = _cov_without_join(r, lane_scopes, cc_only_scopes)
+        cov, flag = verdict or _joined_cov(r, ambiguous, coverage_by_path, start_index)
+        scored.append(_finish(r, cov, flag, target=target, scope_targets=scope_targets,
+                              shared_span=_on_shared_span(r, verdict, shared)))
     return scored

{crapkit-0.7.4 → crapkit-0.7.6}/src/crapkit/store.py

@@ -70,7 +70,7 @@ _CODE_DDL = """CREATE TABLE IF NOT EXISTS {table} (
 # and a store is a file people copy between machines. A name from outside this
 # list is still stored, at a code minted after these.
 _CODE_SEEDS = {"flags": ("measured", "untested", "no-lane", "cc-only"),
-               "remedies": ("ok", "add-tests", "decompose")}
+               "remedies": ("ok", "add-tests", "decompose", "split-lines")}
 
 _SCHEMA = f"""
 CREATE TABLE IF NOT EXISTS runs (

{crapkit-0.7.4 → crapkit-0.7.6/src/crapkit.egg-info}/PKG-INFO

@@ -1,11 +1,11 @@
 Metadata-Version: 2.4
 Name: crapkit
-Version: 0.7.4
+Version: 0.7.6
 Summary: Scores every function on complexity times uncovered risk, ranks the worst, and blocks commits that add more.
 Author: Jean-Francois Gagne
 License: MIT
 Project-URL: Homepage, https://github.com/JeanFrancoisGagne/crapkit
-Project-URL: Documentation, https://jeanfrancoisgagne.github.io/crapkit/handbook.html
+Project-URL: Documentation, https://www.jfgagne.com/crapkit/handbook.html
 Project-URL: Changelog, https://github.com/JeanFrancoisGagne/crapkit/blob/main/CHANGELOG.md
 Project-URL: Issues, https://github.com/JeanFrancoisGagne/crapkit/issues
 Project-URL: Source, https://github.com/JeanFrancoisGagne/crapkit
@@ -111,6 +111,9 @@ crapkit ratchet seed
 git add crapkit.toml crapkit-ratchet.tsv .gitignore
 ```
 
+Not a Python repo? `uvx crapkit init` runs the same commands and adds nothing to your
+manifest: see [A repo that is not Python](#a-repo-that-is-not-python).
+
 `init` detects pytest, Vitest and Jest from the repository's own files. Review the
 generated config before running its commands. When detection leaves a commented
 lane, fill it in using the [lane recipes](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md).
@@ -182,6 +185,28 @@ pip install git+https://github.com/JeanFrancoisGagne/crapkit.git
 pip install .
 ```
 
+### A repo that is not Python
+
+crapkit is a command-line tool, never a dependency of the code it scores. A TypeScript,
+Go or Rust repo adds nothing to its own manifest. With [uv](https://docs.astral.sh/uv/)
+on the machine, `uvx` fetches crapkit into a cache of its own and runs it:
+
+```
+$ uvx crapkit init
+wrote crapkit.toml with 1 scope(s): src
+detected 1 lane(s) from this repo's own files: js - next: run `crapkit coverage`
+added to .gitignore: .crapkit/
+
+$ uvx crapkit coverage
+$ uvx crapkit worklist
+```
+
+The lane still runs your own test runner, so Vitest or Jest and its coverage package come
+from the repo's `node_modules` as they do today. `uv tool install crapkit` or
+`pipx install crapkit` puts a `crapkit` command on PATH once, which is what the
+[commit gate](#the-gate) and the Claude Code plugin call. uv brings its own Python when
+the machine has none.
+
 Requires Python 3.11 or newer and Git on PATH. The CLI has one runtime dependency,
 `lizard>=1.24.0`; a package mirror needs both distributions. Install into the environment
 you intend to use, then check `crapkit --version`. The `pip install -e ".[dev]"` under
@@ -199,7 +224,7 @@ files. Review those commands before running Crapkit in a repository you do not t
 
 ```
 $ crapkit --version
-crapkit 0.7.4
+crapkit 0.7.6
 ```
 
 `python -m crapkit` works identically to the console script and is what to use from a
@@ -455,7 +480,7 @@ crapkit ships a `.pre-commit-hooks.yaml` declaring `id: crapkit-gate`. In your
 repos:
   - repo: https://github.com/JeanFrancoisGagne/crapkit
     # crapkit's release step rewrites this line to the tag it just cut
-    rev: v0.7.4
+    rev: v0.7.6
     hooks:
       - id: crapkit-gate
 ```
@@ -563,7 +588,7 @@ to a workflow, and every input has a default:
       - uses: actions/checkout@v4
         with:
           fetch-depth: 0
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
 ```
 
 The whole job those four lines sit in:
@@ -583,7 +608,7 @@ jobs:
         with:
           python-version: "3.12"       # the interpreter the install below lands in
       - run: pip install -e ".[dev]"   # whatever your lanes need to run
-      - uses: JeanFrancoisGagne/crapkit@v0.7.4
+      - uses: JeanFrancoisGagne/crapkit@v0.7.6
         with:
           gate: "false"
 ```
@@ -654,7 +679,7 @@ closes it. A verify that passed is one line: `**verify passed.** Run 2 against b
 The rows are the ranked worklist for the files the pull request changed, worst first,
 `top` of them, with the rows a finding names listed first. `risk` is ccn times churn
 weight, the number `crapkit worklist` ranks on, and `remedy` is the run's own verdict for
-that function: `decompose`, `add-tests` or `ok`. `(accepted debt)` marks a function the
+that function: `decompose`, `split-lines`, `add-tests` or `ok`. `(accepted debt)` marks a function the
 committed ratchet carries a mark for, so an untouched `legacy_router` does not read like
 the pull request's own new function. A pull request that touches no ranked function gets
 the heading and no table.
@@ -830,6 +855,7 @@ The coverage summary counts all four as `measured` / `untested` / `no_lane` / `c
 | Remedy | Condition | Action |
 |---|---|---|
 | `decompose` | `ccn > ceiling` | Split it. No amount of coverage clears this. |
+| `split-lines` | `ccn <= ceiling`, `crap > ceiling`, and another function shares its source lines | Put each definition on its own lines, then measure again. Coverage cannot tell functions on one line apart, so the score stays at uncovered whatever the tests do. |
 | `add-tests` | `ccn <= ceiling` and `crap > ceiling` | Cover the branches. |
 | `ok` | `crap <= ceiling` | Nothing. |
 
@@ -1268,12 +1294,12 @@ export function classify(row: Row): string {
 ```
 $ crapkit rescore src/grade.ts --gate
 rescore vs run 1 @ 8bfbe613fcd (coverage STALE, complexity fresh)
-   ccn   cov     crap  remedy     function
-     5    0%     30.0  add-tests  src/grade.ts:22  band ( score )
-     5    0%     30.0  add-tests  src/grade.ts:38  demote ( letter , row Row )
-     4    0%     20.0  add-tests  src/grade.ts:8  penalty ( attempts , late )
-     4   45%      6.7  add-tests  src/grade.ts:48  classify ( row Row )
-     4   75%      4.2  ok         src/grade.ts:59  average ( scores Array )
+   ccn   cov     crap  remedy      function
+     5    0%     30.0  add-tests   src/grade.ts:22  band ( score )
+     5    0%     30.0  add-tests   src/grade.ts:38  demote ( letter , row Row )
+     4    0%     20.0  add-tests   src/grade.ts:8  penalty ( attempts , late )
+     4   45%      6.7  add-tests   src/grade.ts:48  classify ( row Row )
+     4   75%      4.2  ok          src/grade.ts:59  average ( scores Array )
 ```
 
 Exit 0: every piece is at or under 6. The `crap` column is loud because its coverage half
@@ -1340,7 +1366,7 @@ with no debt.
 
 | Page | Covers |
 |---|---|
-| [The handbook](https://jeanfrancoisgagne.github.io/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://jeanfrancoisgagne.github.io/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
+| [The handbook](https://www.jfgagne.com/crapkit/handbook.html) | **Start here for anything deeper.** The illustrated handbook: what crapkit is, how every piece works, and where each command earns its keep. Also at [docs/handbook.html](https://www.jfgagne.com/crapkit/handbook.html), self-contained, so it opens straight from a clone. |
 | [docs/adoption.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/adoption.md) | The judgment layer over the quickstarts: scope granularity, exclude vs lane, scoped_tests wiring, the first-verify taint hazard. |
 | [docs/configuration.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/configuration.md) | Every `crapkit.toml` key: type, default, and what it does. |
 | [docs/lanes.md](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/lanes.md) | The lane model, vitest and jest and pytest recipes, artifact reuse, flake retest, containers. |
@@ -1371,6 +1397,12 @@ line arms the complexity gate. See
 for development and [the verified implementation report](https://github.com/JeanFrancoisGagne/crapkit/blob/main/docs/architecture/2026-09-07-implementation/REPORT.md)
 for complete Windows source and Linux wheel results, focused benchmarks and their limits.
 
+## Maintainer and project background
+
+crapkit is created and maintained by [Jean-François Gagné](https://www.jfgagne.com/).
+Read the [project background](https://www.jfgagne.com/projects/crapkit/) for the
+problem it addresses and how it fits into his work on software and AI.
+
 ## License
 
 MIT. See [LICENSE](https://github.com/JeanFrancoisGagne/crapkit/blob/main/LICENSE).