CB-WP-0019 T01/T02: AM-4b asks what a contributor acquires

The two AM-4 budgets had the SAME scope -- one package, no dev edges --
while claiming to bound different things. AM-4b now measures the
workspace with dev edges: 57 crates / 725,258 lines where it read 29 /
317,021, having been blind to 28 crates and 408,237 lines, more source
than its own target.

Target 745,000, ~2.7% of room -- the same margin ADR-0008 D3 gave AM-4a,
applied to a number that grew because the instrument was repaired, not
because anything was added. The target moved to fit the measurement.

T02: proc-macros are COUNTED here and excluded from AM-4a, on purpose.
AM-4a asks what ships and a proc-macro never ships. AM-4b asks what is
acquired, and ADR-0007 D3's acquisition rule counts what the build
fetches -- 'it does not ship' is no answer to 'we downloaded it'. When
the rules disagree, the question each budget asks decides. Measured
share 109,585 lines / 15.1% against AM-4a's 36.2%, so ADR-0008 D2's
refusal to borrow the ratio was right by more than a factor of two.

Caught by this project's own earlier work twice: the mutation
find-string went stale and --self-test reported it BUILD-FREE (the check
CB-WP-0015 added after AM-4a's rotted for two passes), then the DFD gate
caught facts.toml carrying the old numbers.

CB-EV-0001 and ADR-0004 carried live fact: tags on historical readings.
A dated record asserting a CURRENT value is a category error, so those
occurrences are marked as-measured instead of retro-edited, and ADR-0004
gains a supersession note.

make all exits 0.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

tools/dep-weight.py

@@ -27,19 +27,39 @@ import sys
 
 from repo import cargo_bin, enter_root
 
+# AM-4a asks what a GAME SHIPS, so it measures one package's non-dev
+# graph. AM-4b asks what a CONTRIBUTOR ACQUIRES, so it measures the whole
+# workspace including dev edges (CB-WP-0019 T01). They had the same scope
+# until 2026-08-03, which left AM-4b blind to 28 crates and 408,237 lines
+# — more source than its own target.
 PACKAGE = "games-ground"
+WORKSPACE = "--workspace"
 # ADR-0008 D2. `--edges normal` includes proc-macro crates, which run in
 # the compiler and never reach a shipped binary — 89,048 lines, 36.2% of
 # what this tool used to call "what a game ships", `syn` alone 66,916. The
 # shipped-runtime configuration now excludes them.
 #
-# AM-4b is deliberately NOT corrected here: its proc-macro share has not
-# been measured, and correcting a second instrument on the strength of the
-# first one's ratio is the error this change exists to fix.
+# AM-4b's proc-macro share is now MEASURED: 109,585 lines, 15.1% of its
+# real graph (CB-WP-0019 T02). It is deliberately **not** excluded, and
+# that is the opposite of AM-4a's treatment for a stated reason:
+#
+#   AM-4a excludes proc-macros because they run in the compiler and never
+#   reach a shipped binary — counting them in "what a game ships" was
+#   simply false.
+#
+#   AM-4b counts them, because ADR-0007 D3's acquisition rule counts what
+#   the build causes to be FETCHED, and a proc-macro is fetched, compiled
+#   and unaudited on a contributor's machine exactly like any other
+#   dependency. "It does not ship" is no answer to "we downloaded it".
+#
+# When the two rules disagree, the question each budget asks decides:
+# AM-4a asks what ships, AM-4b asks what is acquired.
+#
+# ADR-0008 D2 said this share was unmeasured. That is no longer true.
 PROC_MACRO_EXCLUDED = ["--edges", "normal,no-proc-macro"]
 CONFIGS = {
-    "shipped-runtime": ["--no-default-features"] + PROC_MACRO_EXCLUDED,
-    "dev-toolchain": [],
+    "shipped-runtime": ["-p", PACKAGE, "--no-default-features"] + PROC_MACRO_EXCLUDED,
+    "dev-toolchain": [WORKSPACE, "--edges", "normal,dev"],
 }
 
 # AM-4a / AM-4b targets from specs/GameKernel.md §4. Breaching one fails
@@ -50,9 +70,19 @@ CONFIGS = {
 # 161,000 keeps ~2.4% of room where 250,000 kept ~1.5% — the small
 # rounding up is the only thing this decision gives back, because a target
 # with 1.5% of room fails on a dependency's patch release.
+# CB-WP-0019 T01: the target moves to fit the measurement, never the
+# reverse. AM-4b now measures 725,258 where it used to read 317,021 — not
+# because anything was added, but because it started looking at what it
+# always claimed to bound. A 350,000 target against a 725,258 reading
+# would be a budget that is simply breached, which teaches nothing.
+#
+# 745,000 keeps ~2.7% of room, on ADR-0008 D3's reasoning that a target
+# with ~1.5% fails on a dependency's patch release. It is NOT generosity:
+# it is the same margin AM-4a got, applied to a number that grew because
+# the instrument was fixed.
 TARGETS = {
     "shipped-runtime": 161_000,
-    "dev-toolchain": 350_000,
+    "dev-toolchain": 745_000,
 }
 
 
@@ -70,7 +100,10 @@ def crates(extra_args):
         )
         sys.exit(1)
     out = subprocess.run(
-        [cargo, "tree", "-p", PACKAGE, "--prefix", "none"]
+        # The package/workspace selector now comes from the config, so the
+        # two budgets can ask different questions. Before CB-WP-0019 both
+        # were pinned to one package, which is what made AM-4b blind.
+        [cargo, "tree", "--prefix", "none"]
         + (extra_args if "--edges" in extra_args else ["--edges", "normal"] + extra_args),
         capture_output=True,
         text=True,
@@ -109,6 +142,19 @@ def source_lines(name, version):
     return 0
 
 
+def _dev_only_dependency_is_counted():
+    """AM-4b must actually see dev edges — the defect it was blind to.
+
+    `quick-js` is a dev-dependency of `cb-render-html` and is the crate
+    that exposed the scope defect: it landed in CB-WP-0014, AM-4b did not
+    move, and the ADR that added it withdrew its own cost argument as a
+    result. If this budget stops seeing it, the blindness is back.
+    """
+    dev = crates(CONFIGS["dev-toolchain"])
+    shipped = crates(CONFIGS["shipped-runtime"])
+    return "quick-js" in dev, "quick-js" in shipped
+
+
 def self_test():
     """Each assertion pins a failure this tool must detect.
 
@@ -119,6 +165,21 @@ def self_test():
     """
     results = []
 
+    def _check(name, ok, detail=""):
+        results.append((name, ok, detail))
+
+    # CB-WP-0019 T01: the two budgets must ask DIFFERENT questions, and
+    # `quick-js` is the case that proves it. It is a dev-dependency of
+    # cb-render-html; it landed in CB-WP-0014, AM-4b did not move, and
+    # ADR-0009 withdrew its own cost argument as a result. If AM-4b stops
+    # seeing it the blindness is back; if AM-4a starts seeing it, the
+    # shipped budget has been widened by accident.
+    in_dev, in_shipped = _dev_only_dependency_is_counted()
+    _check("AM-4b sees a dev-only dependency", in_dev,
+           "quick-js is dev-only and is what exposed the scope defect")
+    _check("AM-4a does NOT see a dev-only dependency", not in_shipped,
+           "the shipped budget must stay about what ships")
+
     def check(name, ok, detail=""):
         results.append((name, ok, detail))
 

tools/mutation-check.py

@@ -135,10 +135,15 @@ def rows():
             mutate=("tools/dep-weight.py",
                     '"shipped-runtime": 161_000,', '"shipped-runtime": 1_000,')),
 
-        Row("AM-4b", "third-party LOC, dev toolchain <= 350,000",
+        # CB-WP-0019 T01 widened this to the whole workspace with dev
+        # edges, and the literal moved with it. The stale find-string was
+        # caught build-free by `--self-test`, which is the check
+        # CB-WP-0015 added after AM-4a's mutation rotted unnoticed for two
+        # passes. Second catch, first one that cost nothing.
+        Row("AM-4b", "third-party LOC, what a contributor acquires <= 745,000",
             verify=py + ["tools/dep-weight.py"],
             mutate=("tools/dep-weight.py",
-                    '"dev-toolchain": 350_000,', '"dev-toolchain": 1_000,')),
+                    '"dev-toolchain": 745_000,', '"dev-toolchain": 1_000,')),
 
         # Deliberately RETAINED in this denominator after its withdrawal
         # from the acceptance table. Dropping it would move M-D1-MUT from