diff --git a/CLAUDE.md b/CLAUDE.md
index 30a57dd50e..a26a9bd656 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -8,7 +8,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 Perry is a native TypeScript compiler written in Rust that compiles TypeScript source code directly to native executables. It uses SWC for TypeScript parsing and LLVM for code generation.
 
-**Current Version:** 0.5.1432
+**Current Version:** 0.5.1433
 
 
 ## TypeScript Parity Status
diff --git a/Cargo.lock b/Cargo.lock
index f1406f4a65..d6931f6a61 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -5547,7 +5547,7 @@ checksum = "9b4f627cb1b25917193a259e49bdad08f671f8d9708acfd5fe0a8c1455d87220"
 
 [[package]]
 name = "perry"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "base64",
@@ -5607,14 +5607,14 @@ dependencies = [
 
 [[package]]
 name = "perry-api-manifest"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "serde",
 ]
 
 [[package]]
 name = "perry-audio-miniaudio"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "cc",
  "libc",
@@ -5622,7 +5622,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "inkwell",
@@ -5639,7 +5639,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-arkts"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-hir",
@@ -5647,7 +5647,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-glance"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-hir",
@@ -5655,7 +5655,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-js"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-dispatch",
@@ -5664,7 +5664,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-swiftui"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-hir",
@@ -5672,7 +5672,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-wasm"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "base64",
@@ -5684,7 +5684,7 @@ dependencies = [
 
 [[package]]
 name = "perry-codegen-wear-tiles"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-hir",
@@ -5692,7 +5692,7 @@ dependencies = [
 
 [[package]]
 name = "perry-container-compose"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -5721,14 +5721,14 @@ dependencies = [
 
 [[package]]
 name = "perry-container-e2e"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
 ]
 
 [[package]]
 name = "perry-diagnostics"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "serde",
  "serde_json",
@@ -5736,7 +5736,7 @@ dependencies = [
 
 [[package]]
 name = "perry-dispatch"
-version = "0.5.1432"
+version = "0.5.1433"
 
 [[package]]
 name = "perry-doc-fixture-my-bindings"
@@ -5747,7 +5747,7 @@ dependencies = [
 
 [[package]]
 name = "perry-doc-tests"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "clap",
@@ -5762,7 +5762,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-ads"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "block2",
  "objc2",
@@ -5772,7 +5772,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-argon2"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "argon2",
  "perry-ffi",
@@ -5780,7 +5780,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-axios"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "reqwest",
@@ -5789,7 +5789,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-bcrypt"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bcrypt",
  "perry-ffi",
@@ -5797,7 +5797,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-better-sqlite3"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "rusqlite",
@@ -5805,7 +5805,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-cheerio"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "scraper",
@@ -5813,7 +5813,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-commander"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "perry-runtime",
@@ -5821,7 +5821,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-cron"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "chrono",
  "cron",
@@ -5831,7 +5831,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-dayjs"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "chrono",
  "perry-ffi",
@@ -5839,7 +5839,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-decimal"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "rust_decimal",
@@ -5847,7 +5847,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-dotenv"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "serde_json",
@@ -5855,7 +5855,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-ethers"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "rand 0.10.1",
@@ -5863,7 +5863,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-events"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "perry-runtime",
@@ -5871,14 +5871,14 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-exponential-backoff"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
 ]
 
 [[package]]
 name = "perry-ext-fastify"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bytes",
  "http-body-util",
@@ -5896,7 +5896,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-fetch"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bytes",
  "lazy_static",
@@ -5909,7 +5909,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-http"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bytes",
  "h2",
@@ -5933,7 +5933,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-ioredis"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "lazy_static",
  "perry-ffi",
@@ -5943,7 +5943,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-jsonwebtoken"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "jsonwebtoken",
@@ -5954,7 +5954,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-lru-cache"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "lru",
  "perry-ffi",
@@ -5963,7 +5963,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-moment"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "chrono",
  "perry-ffi",
@@ -5971,7 +5971,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-mongodb"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bson",
  "futures-util",
@@ -5983,7 +5983,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-mysql2"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "chrono",
  "perry-ffi",
@@ -5993,7 +5993,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-nanoid"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "nanoid",
  "perry-ffi",
@@ -6002,7 +6002,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-net"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "bytes",
  "perry-ffi",
@@ -6015,7 +6015,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-node-forge"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "const-oid 0.9.6",
  "der 0.7.10",
@@ -6034,7 +6034,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-nodemailer"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "lettre",
  "perry-ffi",
@@ -6044,7 +6044,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-pdf"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "printpdf",
@@ -6052,7 +6052,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-pg"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "sqlx",
@@ -6061,7 +6061,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-ratelimit"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "governor",
  "perry-ffi",
@@ -6069,7 +6069,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-sharp"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "fast_image_resize",
  "image",
@@ -6079,14 +6079,14 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-slugify"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
 ]
 
 [[package]]
 name = "perry-ext-streams"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "lazy_static",
  "perry-ffi",
@@ -6095,7 +6095,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-undici"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "perry-runtime",
@@ -6104,7 +6104,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-uuid"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "uuid",
@@ -6112,7 +6112,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-validator"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ffi",
  "regex",
@@ -6122,7 +6122,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-ws"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "futures-util",
  "lazy_static",
@@ -6135,7 +6135,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ext-zlib"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "brotli",
  "flate2",
@@ -6145,7 +6145,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ffi"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "dashmap",
  "once_cell",
@@ -6154,7 +6154,7 @@ dependencies = [
 
 [[package]]
 name = "perry-hir"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-api-manifest",
@@ -6172,7 +6172,7 @@ dependencies = [
 
 [[package]]
 name = "perry-parser"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-diagnostics",
@@ -6184,7 +6184,7 @@ dependencies = [
 
 [[package]]
 name = "perry-runtime"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "base64",
@@ -6226,14 +6226,14 @@ dependencies = [
 
 [[package]]
 name = "perry-runtime-static"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-runtime",
 ]
 
 [[package]]
 name = "perry-stdlib"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "aes 0.8.4",
  "aes 0.9.1",
@@ -6328,14 +6328,14 @@ dependencies = [
 
 [[package]]
 name = "perry-stdlib-static"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-stdlib",
 ]
 
 [[package]]
 name = "perry-transform"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "perry-hir",
@@ -6344,14 +6344,14 @@ dependencies = [
 
 [[package]]
 name = "perry-ui"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ui-model",
 ]
 
 [[package]]
 name = "perry-ui-android"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "itoa",
@@ -6368,7 +6368,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-geisterhand"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "rand 0.10.1",
  "serde",
@@ -6378,7 +6378,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-gtk4"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "cairo-rs 0.22.0",
@@ -6401,7 +6401,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-ios"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "block2",
@@ -6417,7 +6417,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-macos"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "block2",
@@ -6432,7 +6432,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-model"
-version = "0.5.1432"
+version = "0.5.1433"
 
 [[package]]
 name = "perry-ui-test"
@@ -6443,11 +6443,11 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-testkit"
-version = "0.5.1432"
+version = "0.5.1433"
 
 [[package]]
 name = "perry-ui-tvos"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "block2",
@@ -6463,7 +6463,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-visionos"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "block2",
@@ -6479,7 +6479,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-watchos"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "block2",
  "libc",
@@ -6492,7 +6492,7 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-windows"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "base64",
  "libc",
@@ -6509,14 +6509,14 @@ dependencies = [
 
 [[package]]
 name = "perry-ui-windows-winui"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "perry-ui-windows",
 ]
 
 [[package]]
 name = "perry-updater"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "anyhow",
  "base64",
@@ -6532,7 +6532,7 @@ dependencies = [
 
 [[package]]
 name = "perry-wasm-host"
-version = "0.5.1432"
+version = "0.5.1433"
 dependencies = [
  "wasmi",
 ]
diff --git a/Cargo.toml b/Cargo.toml
index 319169ecdb..af641b50d0 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -315,7 +315,7 @@ codegen-units = 16
 codegen-units = 16
 
 [workspace.package]
-version = "0.5.1432"
+version = "0.5.1433"
 edition = "2021"
 license = "MIT"
 repository = "https://github.com/PerryTS/perry"
diff --git a/changelog.d/7729-gc-zeal-allocation-pacing.md b/changelog.d/7729-gc-zeal-allocation-pacing.md
index b7726166d2..ed5843b05d 100644
--- a/changelog.d/7729-gc-zeal-allocation-pacing.md
+++ b/changelog.d/7729-gc-zeal-allocation-pacing.md
@@ -8,7 +8,7 @@
 
   The earlier build that ran "instantly with the correct answer" was **vacuous**. `PERRY_GC_MOVING_LOOP_POLLS` was default-OFF there (#7161), so a compute-only program reached no loop safepoint and zeal forced nothing; that build also predates the #7604 exit verdict, so it exited 0 in silence. #7721 flipped the poll default ON — correctly, it is a large collector win — and in the same commit turned zeal from free-and-vacuous into correct-but-unusable. Isolated rather than assumed: the *old* compiler with `PERRY_GC_MOVING_LOOP_POLLS=1` forced at compile and run time already costs 35.8 s for one round under zeal against 0.62 s without, so no commit broke zeal — zeal was never paced, and the poll default is what exposed it. #7254 had already logged "a striking concentration of multi-minute-plus runs" under this pairing and left the population untriaged; this is that triage.
 
-  Zeal now forces a collection at the first poll at which `PERRY_GC_ZEAL_ALLOC_KB` (default 4) of new nursery material has accumulated — the model V8 (`--gc-interval`) and SpiderMonkey (`gcZeal(mode, frequency)`) both use, and for the same reason. The stride is a **monotone high-water mark**, not a "bytes since" delta: each forced collection rearms to `from_space_after + stride`, so a collection that reclaims nothing (an escalation to a non-moving full mark-sweep, which #7592 and #7682 both produced in the field) still demands another full stride of genuinely new allocation. Total forced collections are bounded by `bytes_allocated / stride` whatever the collector does with them, which makes this a bound rather than a hope.
+  Zeal now forces a collection at the first poll at which `PERRY_GC_ZEAL_ALLOC_KB` (default 4) of new nursery material has accumulated — the model V8 (`--gc-interval`) and SpiderMonkey (`gcZeal(mode, frequency)`) both use, and for the same reason. The stride is a **monotone high-water mark**, not a "bytes since" delta: each forced collection rearms to `from_space_after + stride`, so a collection that reclaims nothing (an escalation to a non-moving full mark-sweep, which #7592 and #7682 both produced in the field) still demands another full stride of genuinely new allocation. For any positive stride, total forced collections are bounded by `bytes_allocated / stride` whatever the collector does with them, which makes this a bound rather than a hope. (`PERRY_GC_ZEAL_ALLOC_KB=0` is a supported mode and deliberately outside that bound — it *is* the unpaced every-poll semantics, kept for a bug window that executes exactly once.)
 
   **Measured on the pinned quiet host.** The full workload under zeal goes from a 240 s timeout (unpaced cost ~1,426 s) to **98.8 s with the correct answer**, forcing 193,087 collections out of 2,838,560 polls — all of them copying minors, relocating 3,115,719 objects. The stride sweep below is one binary and one env var, at a quarter scale, and `loop_polls` is **283,852 in every row**, so the knob changes only the decision to collect:
 
@@ -20,7 +20,7 @@
   | 16 | 5,070 | 129,959 | 3.0 s |
   | 64 | 1,291 | 52,357 | 1.1 s |
 
-  Row 0 reproduces the pre-fix 1:1 behaviour exactly on the shipped binary. Every row keeps `copying_minors == forced_collections` and `moved > 0`, so no stride degrades the instrument into non-moving sweeps. 4 KB rather than the faster 16/64 is deliberate — this is a correctness instrument, so the default errs toward sensitivity, still collecting once per ~15 loop iterations while being 14x cheaper than unpaced. The zeal-OFF path is untouched: the same workload without zeal is 4.49 s before and after.
+  Row 0 reproduces the pre-fix behaviour on the shipped binary: 283,857 forced collections for 283,852 polls — one per back-edge poll, plus a handful from the other safepoint zeal forces at (the outermost microtask-pump boundary, which the poll counter does not count). Near 1:1, not exactly. Every row keeps `copying_minors == forced_collections` and `moved > 0`, so no stride degrades the instrument into non-moving sweeps. 4 KB rather than the faster 16/64 is deliberate — this is a correctness instrument, so the default errs toward sensitivity, still collecting once per ~15 loop iterations while being 14x cheaper than unpaced. The zeal-OFF path is untouched: the same workload without zeal is 4.49 s before and after.
 
   `PERRY_GC_ZEAL_ALLOC_KB=0` restores the literal every-poll semantics — the right setting for a small fixture or a bug window executed exactly once. What pacing gives up, stated rather than buried: a window crossed a single time may now fall between two forced collections; a window that *recurs* (every shape in the #7154 family, which is why the reproducers are loops) is still caught, after N KB of allocation instead of on the first iteration.
 
diff --git a/changelog.d/7739-gc-pacing-snapshot-boundary.md b/changelog.d/7739-gc-pacing-snapshot-boundary.md
new file mode 100644
index 0000000000..77a1891577
--- /dev/null
+++ b/changelog.d/7739-gc-pacing-snapshot-boundary.md
@@ -0,0 +1,32 @@
+### Fixed
+
+- **The GC trace's `major_pacing` block reports the escalation boundary the collector actually decides on (#7733 review follow-up).**
+
+  `major_pacing_snapshot` recomputed the boundary as `baseline × growth` and dropped the floor on the floor of the function — `let (_floor, growth_num) = major_pacing_config();` — while the predicate it mirrors, `arena_growth_full_escalation_due`, **also** rejects every reading below that same floor. Wherever the floor dominated, the trace named a boundary the collector does not use:
+
+  | state | predicate escalates at | old snapshot reported |
+  |---|--:|--:|
+  | no full yet (`baseline == 0`) | 32 MB (the floor) | `0` — "escalates at any size" |
+  | `baseline = 4 MB`, growth 2 | 32 MB (the floor) | 8 MB |
+  | `baseline = 64 MB`, growth 2 | 128 MB + 1 | 128 MB |
+
+  That matters more than the size of the diff because of *why* the snapshot exists. #7733 added it so the pacing subject could be asserted **live** in the trace rather than a gate merely proving nothing threw — which is the one job it could not do while misreporting the quantity. This repo has paid for that shape repeatedly (`PERRY_GC_FORCE_EVACUATE` inert for every `gc()`-driven test, #6942/#6946; the matrix's `--pressure` knob disabling the path it was measuring, #7024; `moved=` summing two different collectors, #7025).
+
+  The fix is structural rather than a second correct formula. There is now **one** definition of the boundary, `major_pacing_escalation_threshold_bytes`: `arena_growth_full_escalation_due_inner` is literally `in_use >= it`, and the snapshot reports it verbatim, floor included. `None` means "no arena reading escalates" — either pacing is disabled (`PERRY_GC_MAJOR_PACING_FLOOR_MB=0`) or the growth term overflowed `usize`, which is the same statement about the world; the helper uses `checked_*` rather than `saturating_*` because saturating would report `usize::MAX` and then claim an arena of `usize::MAX` escalates, which the strict `>` clause never would.
+
+  The trace key follows the semantics: **`escalate_at_or_above_bytes`**, replacing `escalate_above_bytes`. The predicate's floor clause is a `>=`, and the old name was half of why the reported figure and the decision could disagree. `null` now means pacing is off. Nothing in `scripts/`, `.github/` or `docs/` consumed the old key.
+
+- **`zeal_holds_the_poll_word_armed_with_nothing_pending` asserts the release, not just the acquire.**
+
+  The test checked `PERRY_GC_POLL_ARMED > 0` inside the guard scope and then only *narrated* the release in a trailing comment. If `ZealGuard`'s `Drop` ever stopped giving the arm back, the process-global word would stay non-zero for the life of the test binary, every later test would silently take the poll's slow path — and this test would still have passed. It now captures the baseline before the guard and asserts `base + 1` inside and `base` after the drop, the shape `a_deferral_arms_the_poll_word_and_draining_disarms_it` already uses a few lines above.
+
+### Docs
+
+- `docs/src/internals/memory-model.md` no longer contradicts itself about back-edge polls: one line said they became default-on in #7721, and a caveat twenty lines below still said "default off since #7161". The caveat now states the current default and its kill switch, the two gaps that survive it (an alloc-free loop body emits no poll by design, `loop_purity::loop_may_allocate`; the specialized `for` / `for-of` / `for-in` lowerings emit none by omission), and the #7604 exit-70 verdict that makes a vacuous zeal run red instead of green.
+- `changelog.d/7729-gc-zeal-allocation-pacing.md` and the matching `gc/zeal.rs` doc drop two overclaims. The `bytes_allocated / stride` bound is qualified to a **positive** stride, since `PERRY_GC_ZEAL_ALLOC_KB=0` is a supported every-poll mode and deliberately outside it. And "reproduces the pre-fix 1:1 behaviour exactly", against a table showing 283,857 collections for 283,852 polls, becomes what the numbers say: one per back-edge poll, plus a handful from the other safepoint zeal forces at — the outermost microtask-pump boundary, which calls `gc_safepoint_moving_minor` without `note_loop_poll_reached`. Near 1:1, not exactly.
+
+### Tests
+
+- `the_reported_escalation_boundary_is_the_one_the_predicate_decides_on` — the named floor-dominates and growth-dominates cases the review called for, plus baseline-zero, the backoff shift and pacing-disabled, then exhaustive over `floor × growth × baseline × shift` probing each boundary's own ±1 neighbourhood. The oracle is a deliberate **independent transcription** of the four clauses the predicate used to spell out inline, not a call into the code under test, so collapsing both onto one helper cannot quietly redefine the rule.
+- `the_shipped_predicate_and_the_shipped_snapshot_read_one_boundary` — drives the real `arena_growth_full_escalation_due` against the real `major_pacing_snapshot` on the live arena, so a future re-split fails here even if the pure helper stays correct. `baseline = 0` is the discriminating row and needs no particular heap size: the old snapshot reported `0` there, i.e. `in_use >= 0`, true of every arena including an empty one, while the predicate declines under the floor.
+- Both are sabotage-checked against the pre-fix formula, and the `ZealGuard` assertion against a neutered `Drop`.
diff --git a/crates/perry-runtime/src/gc/policy.rs b/crates/perry-runtime/src/gc/policy.rs
index 3b8bce56fc..d00d8dca9d 100644
--- a/crates/perry-runtime/src/gc/policy.rs
+++ b/crates/perry-runtime/src/gc/policy.rs
@@ -1593,18 +1593,25 @@ pub(super) fn major_pacing_backoff_shift() -> u32 {
     GC_MAJOR_PACING_BACKOFF_SHIFT.with(|shift| shift.get())
 }
 
-/// `(post-full baseline bytes, backoff shift, arena in-use bytes needed to
-/// escalate the next minor to a full)` — emitted in the GC trace so a gate can
-/// prove the backoff actually engaged rather than merely that nothing threw.
-/// A zero baseline means no full has run yet, and the threshold is reported as
-/// 0 because the `baseline == 0` clause escalates unconditionally.
-#[cfg(feature = "diagnostics")]
-pub(super) fn major_pacing_snapshot() -> (usize, u32, usize) {
-    let (_floor, growth_num) = major_pacing_config();
+/// `(post-full baseline bytes, backoff shift, arena in-use bytes at or above
+/// which the next minor escalates to a full)` — emitted in the GC trace so a
+/// gate can prove the backoff actually engaged rather than merely that nothing
+/// threw.
+///
+/// The third element is [`major_pacing_escalation_threshold_bytes`] verbatim,
+/// i.e. the boundary `arena_growth_full_escalation_due` actually tests, **floor
+/// included**. `None` means arena-growth pacing is disabled
+/// (`PERRY_GC_MAJOR_PACING_FLOOR_MB=0`) and no reading escalates; a zero
+/// baseline (no full yet) reports the floor, which is the reading that
+/// escalates — not `0`, which is what it used to say.
+// `test` as well as `diagnostics` (matching `major_pacing_backoff_shift`), so
+// the test that pins snapshot-vs-predicate agreement still builds under
+// `--no-default-features`, where the trace itself is compiled out.
+#[cfg(any(feature = "diagnostics", test))]
+pub(super) fn major_pacing_snapshot() -> (usize, u32, Option<usize>) {
     let baseline = GC_LAST_FULL_ARENA_IN_USE_BYTES.with(|bytes| bytes.get());
     let shift = major_pacing_backoff_shift();
-    let threshold = baseline.saturating_mul(growth_num.saturating_mul(1usize << shift));
-    (baseline, shift, threshold)
+    (baseline, shift, major_pacing_escalation_threshold_bytes())
 }
 
 #[cfg(test)]
@@ -2645,7 +2652,7 @@ pub(super) fn test_start_budgeted_minor_fallback_state_with_trace(
 /// parse / String alloc. The env vars are for tuning and measurement (read at
 /// process start); defaults chosen so churn oscillates ~baseline..2×baseline
 /// and stays below node's peak.
-fn major_pacing_config() -> (usize, usize) {
+pub(super) fn major_pacing_config() -> (usize, usize) {
     use std::sync::OnceLock;
     static CONFIG: OnceLock<(usize, usize)> = OnceLock::new();
     let &(floor_bytes, growth_num) = CONFIG.get_or_init(|| {
@@ -2685,25 +2692,70 @@ pub(super) fn arena_growth_full_escalation_due() -> bool {
 }
 
 fn arena_growth_full_escalation_due_inner() -> bool {
-    let (floor_bytes, growth_num) = major_pacing_config();
-    if floor_bytes == 0 {
-        return false; // PERRY_GC_MAJOR_PACING_FLOOR_MB=0 disables the pacing
-    }
-    let in_use = crate::arena::arena_in_use_bytes();
-    if in_use < floor_bytes {
-        return false;
+    match major_pacing_escalation_threshold_bytes() {
+        // `PERRY_GC_MAJOR_PACING_FLOOR_MB=0` disables the pacing: no reading
+        // escalates, so there is no boundary to compare against.
+        None => false,
+        Some(threshold) => crate::arena::arena_in_use_bytes() >= threshold,
     }
+}
+
+/// The smallest `arena_in_use_bytes()` reading that escalates the next minor to
+/// a full, or `None` when arena-growth pacing is disabled outright and no
+/// reading escalates.
+///
+/// **This is the only definition of that boundary.**
+/// `arena_growth_full_escalation_due_inner` is now literally
+/// `in_use >= this`, and `major_pacing_snapshot` reports exactly this — so the
+/// decision and the diagnostic cannot drift apart, rather than merely agreeing
+/// today. #7733 added the snapshot so the pacing subject could be asserted live
+/// in the GC trace, and it recomputed the boundary as `baseline × growth` with
+/// the floor dropped on the ground (`let (_floor, growth_num) = …`). Wherever
+/// the floor dominates — the entire pre-first-full phase, where the old
+/// snapshot reported `0`, and any baseline below `floor / growth` — the trace
+/// named a boundary the collector does not use. A probe that misreports the
+/// quantity it exists to prove is this repo's most expensive recurring bug, so
+/// the fix is one source of truth, not two that match.
+fn major_pacing_escalation_threshold_bytes() -> Option<usize> {
+    let (floor_bytes, growth_num) = major_pacing_config();
     let baseline = GC_LAST_FULL_ARENA_IN_USE_BYTES.with(|bytes| bytes.get());
-    // No full yet (baseline 0): bound the initial growth once we clear the floor.
-    if baseline == 0 {
-        return true;
-    }
     // Yield-adaptive: a full that reclaimed almost nothing pushes the next
     // escalation out (`GC_MAJOR_PACING_BACKOFF_SHIFT`). Shift the multiplier,
     // not the baseline, so one productive full restores the original pacing.
     let shift = GC_MAJOR_PACING_BACKOFF_SHIFT.with(|shift| shift.get());
+    major_pacing_escalation_threshold_for(floor_bytes, growth_num, baseline, shift)
+}
+
+/// Pure `(config, state) → boundary`, factored out so the floor/growth
+/// interaction is unit-testable without a 32 MB live heap (which is what left
+/// the original divergence untested: every reachable unit test sat below the
+/// floor, where the two formulas' disagreement is invisible to a `bool`).
+///
+/// `None` means "no arena reading escalates" — either pacing is off, or the
+/// growth term overflowed `usize`, which is the same statement about the world.
+pub(super) fn major_pacing_escalation_threshold_for(
+    floor_bytes: usize,
+    growth_num: usize,
+    baseline: usize,
+    shift: u32,
+) -> Option<usize> {
+    if floor_bytes == 0 {
+        return None; // PERRY_GC_MAJOR_PACING_FLOOR_MB=0 disables the pacing
+    }
+    // No full yet (baseline 0): bound the initial growth once we clear the
+    // floor, so the floor alone is the boundary.
+    if baseline == 0 {
+        return Some(floor_bytes);
+    }
     let growth = growth_num.saturating_mul(1usize << shift);
-    in_use > baseline.saturating_mul(growth)
+    // `+1` because the growth clause is a strict `>` while the floor clause is
+    // a `>=`; taking the max of the two is what the snapshot used to omit.
+    // `checked_*` rather than `saturating_*`: a growth term that does not fit
+    // in a `usize` is a boundary no arena reading can reach, which is exactly
+    // what `None` says — saturating would report `usize::MAX` and then claim an
+    // arena of `usize::MAX` escalates, which the `>` clause never would.
+    let growth_boundary = baseline.checked_mul(growth)?.checked_add(1)?;
+    Some(floor_bytes.max(growth_boundary))
 }
 
 fn gc_start_budgeted_cycle_for_pressure(progress_kind: GcProgressKind) -> Option<BudgetedGcCycle> {
diff --git a/crates/perry-runtime/src/gc/telemetry.rs b/crates/perry-runtime/src/gc/telemetry.rs
index 4271d4e54e..525afcb6d4 100644
--- a/crates/perry-runtime/src/gc/telemetry.rs
+++ b/crates/perry-runtime/src/gc/telemetry.rs
@@ -1118,10 +1118,15 @@ impl GcCycleTrace {
         let steps_value = steps_json(self.steps_before, steps_after);
         let (pacing_baseline, pacing_shift, pacing_threshold) =
             super::policy::major_pacing_snapshot();
+        // `escalate_at_or_above_bytes`, not the old `escalate_above_bytes`: the
+        // predicate this mirrors is `in_use >= threshold` (its floor clause is a
+        // `>=`), and the old name was half of why the reported figure and the
+        // decision could disagree. `null` = arena-growth pacing disabled
+        // (`PERRY_GC_MAJOR_PACING_FLOOR_MB=0`), i.e. no reading escalates.
         let major_pacing_json = serde_json::json!({
             "baseline_bytes": pacing_baseline,
             "backoff_shift": pacing_shift,
-            "escalate_above_bytes": pacing_threshold,
+            "escalate_at_or_above_bytes": pacing_threshold,
         });
         serde_json::json!({
             "event": "gc_cycle",
diff --git a/crates/perry-runtime/src/gc/tests/triggers.rs b/crates/perry-runtime/src/gc/tests/triggers.rs
index 7c4cec3270..4b4a9abb11 100644
--- a/crates/perry-runtime/src/gc/tests/triggers.rs
+++ b/crates/perry-runtime/src/gc/tests/triggers.rs
@@ -778,6 +778,158 @@ fn declining_to_escalate_records_no_pre_full_reading() {
     );
 }
 
+/// The escalation boundary the GC trace reports must be the boundary the
+/// predicate takes its decision on.
+///
+/// #7733 added `major_pacing_snapshot` for exactly one purpose: so a gate could
+/// assert the pacing subject was LIVE in the trace rather than merely that
+/// nothing threw. It then recomputed the boundary as `baseline × growth` with
+/// the floor dropped on the floor of the function (`let (_floor, growth_num)`),
+/// while `arena_growth_full_escalation_due` rejects everything under that same
+/// floor. Wherever the floor dominates the two disagreed — most starkly before
+/// the first full, where the trace reported `0` ("escalates at any size") for a
+/// collector that escalates at 32 MB. A probe that misreports its own subject
+/// is worse than no probe, because it reads green.
+///
+/// `escalates_reference` below is a deliberate INDEPENDENT transcription of the
+/// four clauses the predicate used to spell out inline — not a call into the
+/// code under test. The point is to pin the boundary against a second statement
+/// of the rule, so collapsing both onto one helper cannot quietly redefine it.
+#[test]
+fn the_reported_escalation_boundary_is_the_one_the_predicate_decides_on() {
+    use super::super::policy::major_pacing_escalation_threshold_for;
+    const MB: usize = 1024 * 1024;
+
+    fn escalates_reference(
+        floor: usize,
+        growth_num: usize,
+        baseline: usize,
+        shift: u32,
+        in_use: usize,
+    ) -> bool {
+        if floor == 0 {
+            return false; // pacing disabled
+        }
+        if in_use < floor {
+            return false; // under the absolute floor: small heaps never pay
+        }
+        if baseline == 0 {
+            return true; // no full yet
+        }
+        in_use > baseline.saturating_mul(growth_num.saturating_mul(1usize << shift))
+    }
+
+    // The named cases first: each is a different bug if it breaks.
+    assert_eq!(
+        major_pacing_escalation_threshold_for(32 * MB, 2, 4 * MB, 0),
+        Some(32 * MB),
+        "FLOOR DOMINATES (4 MB baseline × 2 = 8 MB, under the 32 MB floor): the \
+         boundary is the floor. Reporting 8 MB names a collection the collector \
+         will decline to run."
+    );
+    assert_eq!(
+        major_pacing_escalation_threshold_for(32 * MB, 2, 64 * MB, 0),
+        Some(128 * MB + 1),
+        "GROWTH DOMINATES: strictly ABOVE K× the baseline, per the `>` clause"
+    );
+    assert_eq!(
+        major_pacing_escalation_threshold_for(32 * MB, 2, 0, 0),
+        Some(32 * MB),
+        "no full yet: the floor is the boundary — the old snapshot reported 0 \
+         here, i.e. 'always', for a collector that escalates at 32 MB"
+    );
+    assert_eq!(
+        major_pacing_escalation_threshold_for(32 * MB, 2, 64 * MB, 2),
+        Some(512 * MB + 1),
+        "the backoff shift multiplies the GROWTH term, not the floor"
+    );
+    assert_eq!(
+        major_pacing_escalation_threshold_for(0, 2, 64 * MB, 0),
+        None,
+        "PERRY_GC_MAJOR_PACING_FLOOR_MB=0 disables the pacing outright: no \
+         arena reading escalates, so there is no boundary to report"
+    );
+
+    // ...then exhaustively against the reference, probing each boundary's own
+    // ±1 neighbourhood so an off-by-one cannot hide between the named cases.
+    for &floor in &[0, 1, 8 * MB, 32 * MB] {
+        for &growth_num in &[1, 2, 3] {
+            for &baseline in &[0, 1, 4 * MB, 32 * MB, usize::MAX / 4, usize::MAX] {
+                for shift in 0..=2u32 {
+                    let threshold =
+                        major_pacing_escalation_threshold_for(floor, growth_num, baseline, shift);
+                    let mut probes =
+                        vec![0usize, 1, 8 * MB, 32 * MB, 64 * MB, 512 * MB, usize::MAX];
+                    if let Some(boundary) = threshold {
+                        probes.extend([
+                            boundary.saturating_sub(1),
+                            boundary,
+                            boundary.saturating_add(1),
+                        ]);
+                    }
+                    for in_use in probes {
+                        assert_eq!(
+                            threshold.is_some_and(|boundary| in_use >= boundary),
+                            escalates_reference(floor, growth_num, baseline, shift, in_use),
+                            "floor={floor} growth={growth_num} baseline={baseline} \
+                             shift={shift} in_use={in_use}: the reported boundary \
+                             disagrees with the escalation rule"
+                        );
+                    }
+                }
+            }
+        }
+    }
+}
+
+/// ...and the SHIPPED pair must agree, not just the pure helper.
+///
+/// The pure test above proves the formula; this proves both production callers
+/// are actually reading it. It drives the real `arena_growth_full_escalation_due`
+/// against the real `major_pacing_snapshot` on this thread's real arena, so a
+/// future re-split (a snapshot that recomputes "the same" boundary inline)
+/// fails here even if the pure helper stays correct.
+///
+/// `baseline = 0` is the discriminating row and it needs no particular heap
+/// size: the pre-#7740 snapshot reported `0` there, i.e. `in_use >= 0`, which
+/// is true of every arena including an empty one, while the predicate declines
+/// under the floor.
+#[test]
+fn the_shipped_predicate_and_the_shipped_snapshot_read_one_boundary() {
+    use super::super::policy::{
+        arena_growth_full_escalation_due, major_pacing_config, major_pacing_snapshot,
+        test_reset_major_pacing_backoff, test_set_major_pacing_baseline,
+    };
+    let (floor_bytes, _growth_num) = major_pacing_config();
+    test_reset_major_pacing_backoff();
+    let saved = test_set_major_pacing_baseline(0);
+
+    for baseline in [0usize, 1, floor_bytes / 4, usize::MAX / 4] {
+        test_set_major_pacing_baseline(baseline);
+        let (reported_baseline, _shift, threshold) = major_pacing_snapshot();
+        let in_use = crate::arena::arena_in_use_bytes();
+        let due = arena_growth_full_escalation_due();
+        // `due == true` records a pre-full reading; drop it so the next row and
+        // every later test on this thread start from a clean pacing state.
+        test_reset_major_pacing_backoff();
+
+        assert_eq!(
+            reported_baseline, baseline,
+            "the snapshot must report the live baseline"
+        );
+        assert_eq!(
+            due,
+            threshold.is_some_and(|boundary| in_use >= boundary),
+            "baseline={baseline}: the trace's escalate_at_or_above_bytes \
+             ({threshold:?}) disagrees with the verdict the collector actually \
+             took on this arena ({in_use} bytes)"
+        );
+    }
+
+    test_set_major_pacing_baseline(saved);
+    test_reset_major_pacing_backoff();
+}
+
 // ---------------------------------------------- the poll's arming word -----
 
 /// The deferral flag and `PERRY_GC_POLL_ARMED` are one piece of state with two
@@ -864,18 +1016,34 @@ fn an_unarmed_poll_touches_nothing() {
 /// path: every back-edge would read zero, skip the call, and force nothing —
 /// and `zeal_liveness_report` would be left to report the vacuity after the
 /// fact instead of the instrument simply working.
+///
+/// The release half is asserted, not narrated. `ZealGuard::set(true)` arms the
+/// process-global word; if its `Drop` ever stopped giving that arm back, the
+/// word would stay non-zero for the life of the test binary and every later
+/// test would silently take the poll's slow path — with this test still green,
+/// because it only ever looked INSIDE the scope. A test that cannot fail is not
+/// a test.
 #[test]
 fn zeal_holds_the_poll_word_armed_with_nothing_pending() {
     let _isolation = GcTestIsolationGuard::new();
     crate::gc::set_safepoint_pending(false);
+    let base = crate::gc::PERRY_GC_POLL_ARMED.load(std::sync::atomic::Ordering::Relaxed);
     {
         let _zeal = super::super::zeal::ZealGuard::set(true);
-        assert!(
-            crate::gc::PERRY_GC_POLL_ARMED.load(std::sync::atomic::Ordering::Relaxed) > 0,
+        assert_eq!(
+            crate::gc::PERRY_GC_POLL_ARMED.load(std::sync::atomic::Ordering::Relaxed),
+            base + 1,
             "zeal must keep the poll reachable even with no deferral outstanding"
         );
     }
     // And it gives the arm back, so one zeal test does not leave every later
     // test in this binary paying for the slow path.
+    assert_eq!(
+        crate::gc::PERRY_GC_POLL_ARMED.load(std::sync::atomic::Ordering::Relaxed),
+        base,
+        "dropping the ZealGuard must release the arm it took — a leaked arm \
+         pins the poll on for the rest of the process, and nothing else in this \
+         binary would notice"
+    );
     crate::gc::set_safepoint_pending(false);
 }
diff --git a/crates/perry-runtime/src/gc/zeal.rs b/crates/perry-runtime/src/gc/zeal.rs
index 849de5176d..a408b74101 100644
--- a/crates/perry-runtime/src/gc/zeal.rs
+++ b/crates/perry-runtime/src/gc/zeal.rs
@@ -210,9 +210,12 @@ pub(crate) fn note_zeal_forced_collection() {
 /// | 16 | 5,070 | 129,959 | 3.0 s |
 /// | 64 | 1,291 | 52,357 | 1.1 s |
 ///
-/// Row 0 is the pre-fix behaviour reproduced exactly — 283,857 collections for
-/// 283,852 polls, i.e. 1:1 — and it is what made the full-scale workload take
-/// ~24 minutes.
+/// Row 0 is the pre-fix behaviour reproduced on the shipped binary: 283,857
+/// collections for 283,852 polls — one per back-edge poll, plus a handful from
+/// the other safepoint zeal forces at (the outermost microtask-pump boundary,
+/// which `note_loop_poll_reached` does not count). Near 1:1, not exactly; the
+/// bound that matters is one per poll, and it is what made the full-scale
+/// workload take ~24 minutes.
 ///
 /// 4 KB rather than the faster 16/64 is deliberate: this is a *correctness*
 /// instrument, so the default errs toward sensitivity. It still collects once
diff --git a/docs/src/internals/memory-model.md b/docs/src/internals/memory-model.md
index 1e602eb4da..deab13496c 100644
--- a/docs/src/internals/memory-model.md
+++ b/docs/src/internals/memory-model.md
@@ -150,10 +150,19 @@ investigation:
   publishes it: the default depth of 4 misses it silently, and
   `PERRY_GC_PROTECT_FROMSPACE_DEPTH=800` faults on the first use. Rule of thumb:
   depth ≥ the number of safepoints the suspect value survives.
-- `PERRY_GC_ZEAL` cannot emit loop back-edge polls that codegen never produced.
-  Those require the **compile-time** `PERRY_GC_MOVING_LOOP_POLLS=1` (default off
-  since #7161). Without it, zeal only fires at event-loop boundaries and a
-  compute-only loop never collects at all. Compile *and* run with the poll opt-in.
+- `PERRY_GC_ZEAL` cannot force a collection at a loop back-edge poll codegen
+  never produced. Those come from the **compile-time**
+  `PERRY_GC_MOVING_LOOP_POLLS`, which is **default ON since #7721** (kill switch
+  `=0`); it was default off from #7161 until then, and that is why zeal used to
+  look free — it was collecting nothing. Two gaps survive the new default:
+  codegen emits no poll for a provably alloc-free loop body (by design,
+  `loop_purity::loop_may_allocate`), nor for the specialized `for` / `for-of` /
+  `for-in` lowerings (by omission). On a poll-free binary zeal fires only at
+  event-loop boundaries, so a compute-only loop never collects at all. You no
+  longer have to remember to check: since #7604 a zeal run prints
+  `[gc-zeal] forced_collections=… copying_minors=… moved_objects=… loop_polls=…`
+  at exit and **exits 70** if it forced or moved nothing, so a vacuous run is
+  red rather than green.
 - **Page protection is Unix-only.** `mprotect` / `sigaction` / `sysconf` are not
   exposed by the `libc` crate on `x86_64-pc-windows-msvc`, a target
   `perry-runtime` is genuinely built for. On non-Unix hosts `=1` degrades to