Repair Cloudflare WordPress corpus dev rig

Author: chubes4

package-lock.json

@@ -82,6 +82,7 @@
     "build": "node ./node_modules/typescript/bin/tsc -b packages/runtime-core packages/runtime-playground packages/cli && node scripts/ensure-cli-bin-executable.mjs",
     "cloudflare:dry-run": "npm exec -- wrangler deploy --dry-run --config packages/runtime-cloudflare/wrangler.jsonc",
     "cloudflare:local-gate": "node scripts/cloudflare-local-gate.mjs",
+    "postinstall": "patch-package",
     "generate:cloudflare-wordpress-runtime-corpus": "tsx scripts/generate-cloudflare-wordpress-runtime-corpus.ts",
     "provision:cloudflare-wordpress-runtime-corpus": "node scripts/provision-cloudflare-wordpress-runtime-corpus.mjs",
     "prepare": "npm run build",
@@ -191,7 +192,6 @@
     "test:php-artifact-import-idempotency": "php scripts/php-artifact-import-idempotency-smoke.php",
     "test:php-browser-runtime-local-package": "php scripts/php-browser-runtime-local-package-smoke.php",
     "test:php-browser-runtime-agent-substrate": "php scripts/php-browser-runtime-agent-substrate-smoke.php",
-    "test:php-browser-preview-only-session": "php -d zend.assertions=1 -d assert.exception=1 scripts/php-browser-preview-only-session-smoke.php",
     "test:mcp-client-configs": "tsx tests/mcp-client-configs.test.ts",
     "test:runtime-tool-policy": "tsx tests/runtime-tool-policy.test.ts",
     "test:primitive-contract-parity": "tsx tests/primitive-contract-parity.test.ts",
@@ -256,6 +256,7 @@
   "devDependencies": {
     "@cloudflare/workers-types": "^5.20260718.1",
     "@types/node": "^24.0.0",
+    "patch-package": "^8.0.1",
     "tsx": "^4.20.0",
     "wrangler": "^4.112.0"
   },
@@ -455,6 +456,7 @@
     "yargs-parser": "^21.1.1"
   },
   "overrides": {
+    "@php-wasm/stream-compression": "3.1.45",
     "ajv": "^8.20.0",
     "express": "^4.22.2",
     "qs": "^6.15.2",

package.json

@@ -0,0 +1,10 @@
+# Temporary Playground Overlay
+
+`@php-wasm+stream-compression+3.1.45.patch` is the built output of
+WordPress Playground PR [#4110](https://github.com/WordPress/wordpress-playground/pull/4110)
+at commit `afea9c042999c65f5b43bc049ca4b2bea9d8b361`.
+
+WP Codebox needs the repaired bounded Range decoder for the Cloudflare dev rig
+and browser asset path before that upstream change can be merged and released.
+Remove this patch and the package override after a published Playground package
+contains the same change.

patches/@php-wasm+stream-compression+3.1.45.patch

@@ -2,6 +2,7 @@ import { spawn } from "node:child_process"
 import { mkdtemp, rm } from "node:fs/promises"
 import { tmpdir } from "node:os"
 import { join } from "node:path"
+import { stripVTControlCharacters } from "node:util"
 
 const port = 8792
 const origin = `http://127.0.0.1:${port}`
@@ -73,7 +74,7 @@ async function stopWorker() {
 async function waitForServer() {
   const deadline = Date.now() + 30_000
   while (Date.now() < deadline) {
-    if (/Ready on http:\/\/(?:localhost|127\.0\.0\.1):8792/.test(output)) return
+    if (/Ready on http:\/\/(?:localhost|127\.0\.0\.1):8792/.test(stripVTControlCharacters(output))) return
     if (child.exitCode !== null) throw new Error(`workerd exited before starting:\n${output}`)
     await new Promise((resolve) => setTimeout(resolve, 100))
   }

patches/README.md

@@ -1,32 +1,23 @@
 import { createHash } from "node:crypto"
 import { mkdir, writeFile } from "node:fs/promises"
 import { dirname, resolve } from "node:path"
-import { decodeRemoteZip, encodeZip } from "@php-wasm/stream-compression"
+import { decodeZip, encodeZip } from "@php-wasm/stream-compression"
 import { isWordPressRuntimeFile } from "../packages/runtime-cloudflare/src/wordpress-runtime-corpus.js"
 import { WORDPRESS_RUNTIME_ARTIFACT_SCHEMA, wordpressRuntimeArtifactKey, type WordPressRuntimeArtifactManifest } from "../packages/runtime-cloudflare/src/wordpress-runtime-artifact.js"
 
 const sourceUrl = process.env.WORDPRESS_RUNTIME_ARCHIVE_URL ?? "https://downloads.wordpress.org/release/wordpress-7.0.2.zip"
 const sourceVersion = process.env.WORDPRESS_RUNTIME_VERSION ?? "7.0.2"
 const output = resolve(process.env.WORDPRESS_RUNTIME_ARTIFACT_OUTPUT ?? "artifacts/cloudflare-wordpress-runtime-corpus.zip")
 const manifestOutput = resolve("packages/runtime-cloudflare/assets/wordpress-runtime-artifact.json")
-const decoder = new TextDecoder()
-const archivePaths = new Set<string>()
-const selected: File[] = []
-
-const response = await fetch(sourceUrl, { method: "HEAD" })
-if (!response.ok) throw new Error(`Unable to inspect WordPress archive: ${response.status}.`)
+const response = await fetch(sourceUrl)
+if (!response.ok || !response.body) throw new Error(`Unable to download WordPress archive: ${response.status}.`)
 const identity = response.headers.get("etag") ?? response.headers.get("last-modified") ?? undefined
-const stream = await decodeRemoteZip(sourceUrl, (entry) => {
-  const path = decoder.decode(entry.path)
-  archivePaths.add(path)
-  return isWordPressRuntimeFile(path, archivePaths)
-})
-for await (const entry of stream) {
-  const path = entry instanceof File ? entry.name : decoder.decode(entry.path)
-  if (!isWordPressRuntimeFile(path, archivePaths)) continue
-  const bytes = entry instanceof File ? new Uint8Array(await entry.arrayBuffer()) : entry.bytes
-  selected.push(new File([bytes], path, { lastModified: 0 }))
-}
+const files: File[] = []
+for await (const file of decodeZip(response.body)) files.push(file)
+const archivePaths = new Set(files.map((file) => file.name))
+const selected = files
+  .filter((file) => isWordPressRuntimeFile(file.name, archivePaths))
+  .map((file) => new File([file], file.name, { lastModified: 0 }))
 selected.sort((left, right) => left.name.localeCompare(right.name))
 if (!selected.length) throw new Error("WordPress archive did not yield a runtime corpus.")
 const archive = new Uint8Array(await new Response(encodeZip(selected)).arrayBuffer())

scripts/cloudflare-local-gate.mjs

@@ -328,7 +328,9 @@ test("WordPress runtime corpus generator keeps the ZIP outside the Worker bundle
   const generator = await readFile(new URL("../scripts/generate-cloudflare-wordpress-runtime-corpus.ts", import.meta.url), "utf8")
   const artifact = await readFile(new URL("../packages/runtime-cloudflare/src/wordpress-runtime-artifact.ts", import.meta.url), "utf8")
   const manifest = JSON.parse(await readFile(new URL("../packages/runtime-cloudflare/assets/wordpress-runtime-artifact.json", import.meta.url), "utf8")) as WordPressRuntimeArtifactManifest
-  assert.match(generator, /decodeRemoteZip\(sourceUrl/)
+  assert.match(generator, /const response = await fetch\(sourceUrl\)/)
+  assert.match(generator, /decodeZip\(response\.body\)/)
+  assert.doesNotMatch(generator, /decodeRemoteZip/)
   assert.match(generator, /encodeZip\(selected\)/)
   assert.match(generator, /lastModified: 0/)
   assert.match(generator, /artifacts\/cloudflare-wordpress-runtime-corpus\.zip/)

scripts/generate-cloudflare-wordpress-runtime-corpus.ts

@@ -1,21 +1,18 @@
 import assert from "node:assert/strict"
 import test from "node:test"
-import { decodeRemoteZip, type CentralDirectoryEntry } from "@php-wasm/stream-compression"
+import { decodeZip } from "@php-wasm/stream-compression"
 import { isWordPressRuntimeFile, summarizeWordPressRuntimeCorpus, type WordPressArchiveEntry } from "../packages/runtime-cloudflare/src/wordpress-runtime-corpus.js"
 
 const WORDPRESS_ARCHIVE_URL = "https://wordpress.org/latest.zip"
 const MAX_RUNTIME_CORPUS_BYTES = 32 * 1024 * 1024
 
 test("WordPress production runtime corpus stays within the Worker materialization budget", async () => {
-  const decoder = new TextDecoder()
   const entries: WordPressArchiveEntry[] = []
-  const stream = await decodeRemoteZip(WORDPRESS_ARCHIVE_URL, (entry) => {
-    const directoryEntry = entry as CentralDirectoryEntry
-    entries.push({ path: decoder.decode(directoryEntry.path), uncompressedSize: directoryEntry.uncompressedSize, isDirectory: directoryEntry.isDirectory })
-    return false
-  })
-  for await (const _ of stream) {
-    // The predicate records central-directory metadata and selects no file ranges.
+  const response = await fetch(WORDPRESS_ARCHIVE_URL)
+  assert.equal(response.ok, true, `WordPress archive returned HTTP ${response.status}.`)
+  assert.ok(response.body, "WordPress archive response has no body.")
+  for await (const file of decodeZip(response.body)) {
+    entries.push({ path: file.name, uncompressedSize: file.size, isDirectory: file.name.endsWith("/") })
   }
 
   const corpus = summarizeWordPressRuntimeCorpus(entries)
@@ -39,5 +36,7 @@ test("WordPress production runtime corpus stays within the Worker materializatio
     legacyBytes: legacy.reduce((total, entry) => total + entry.uncompressedSize, 0),
     largestExcludedGroups: byGroup(excluded),
   }))
+  assert.ok(entries.some((entry) => entry.path === "wordpress/wp-includes/version.php"), "Archive is missing wp-includes/version.php.")
+  assert.ok(corpus.selectedFiles > 0, "WordPress archive did not yield a runtime corpus.")
   assert.ok(corpus.selectedBytes <= MAX_RUNTIME_CORPUS_BYTES, `Selected ${corpus.selectedBytes} bytes exceeds the ${MAX_RUNTIME_CORPUS_BYTES}-byte runtime corpus budget.`)
 })