restructure: move el compiler content into lang/

2026-05-05 01:38:51 -05:00
parent ce68f91a38
commit 1ae68962cf
143 changed files with 0 additions and 0 deletions
@@ -0,0 +1,150 @@
+{
+  "allowlist": "{\"p\":[],\"br\":[],\"strong\":[],\"em\":[],\"u\":[],\"s\":[],\"code\":[],\"pre\":[],\"ul\":[],\"ol\":[],\"li\":[],\"h1\":[],\"h2\":[],\"h3\":[],\"h4\":[],\"blockquote\":[],\"a\":[\"href\",\"title\"]}",
+  "cases": [
+    {
+      "name": "01-pass-through-list",
+      "input":    "<ol><li><strong>x</strong></li></ol>",
+      "expected": "<ol><li><strong>x</strong></li></ol>"
+    },
+    {
+      "name": "02-paragraphs-with-bold",
+      "input":    "<p>hello <strong>world</strong></p><p>second</p>",
+      "expected": "<p>hello <strong>world</strong></p><p>second</p>"
+    },
+    {
+      "name": "03-pre-code-block",
+      "input":    "<pre><code>npm install</code></pre>",
+      "expected": "<pre><code>npm install</code></pre>"
+    },
+    {
+      "name": "04-allowed-https-link",
+      "input":    "<a href=\"https://example.com\">click</a>",
+      "expected": "<a href=\"https://example.com\">click</a>"
+    },
+    {
+      "name": "05-allowed-anchor-link",
+      "input":    "<a href=\"#section\">jump</a>",
+      "expected": "<a href=\"#section\">jump</a>"
+    },
+    {
+      "name": "06-javascript-scheme-blocked",
+      "input":    "<a href=\"javascript:alert(1)\">click</a>",
+      "expected": "<a>click</a>"
+    },
+    {
+      "name": "07-about-scheme-blocked",
+      "input":    "<a href=\"about:blank#x\">click</a>",
+      "expected": "<a>click</a>"
+    },
+    {
+      "name": "08-data-uri-blocked",
+      "input":    "<a href=\"data:text/html,<script>alert(1)</script>\">x</a>",
+      "expected": "<a>x</a>"
+    },
+    {
+      "name": "09-script-content-dropped",
+      "input":    "before<script>alert(1)</script>after",
+      "expected": "beforeafter"
+    },
+    {
+      "name": "10-iframe-content-dropped",
+      "input":    "<iframe src=\"evil\">junk</iframe>safe",
+      "expected": "safe"
+    },
+    {
+      "name": "11-form-content-dropped",
+      "input":    "<form action=\"/steal\">x</form>safe",
+      "expected": "safe"
+    },
+    {
+      "name": "12-img-with-onerror-dropped",
+      "input":    "<img src=x onerror=alert(2)>",
+      "expected": ""
+    },
+    {
+      "name": "13-comment-injection-bypass-blocked",
+      "input":    "<form action=\"-->junk\">x</form>safe",
+      "expected": "safe"
+    },
+    {
+      "name": "14-mixed-legit-and-attack",
+      "input":    "<p>hello</p><script>alert(1)</script><p>world</p>",
+      "expected": "<p>hello</p><p>world</p>"
+    },
+    {
+      "name": "15-pre-encoded-entities-preserved",
+      "input":    "&lt;script&gt;alert(1)&lt;/script&gt;",
+      "expected": "&lt;script&gt;alert(1)&lt;/script&gt;"
+    },
+    {
+      "name": "16-unicode-in-href-preserved",
+      "input":    "<a href=\"https://example.com/?q=日本語\">x</a>",
+      "expected": "<a href=\"https://example.com/?q=日本語\">x</a>"
+    },
+    {
+      "name": "17-unclosed-tag-passes-through",
+      "input":    "<p>unclosed",
+      "expected": "<p>unclosed"
+    },
+    {
+      "name": "18-onclick-attribute-stripped-tag-survives",
+      "input":    "<p onclick=\"x\">hi</p>",
+      "expected": "<p>hi</p>"
+    },
+    {
+      "name": "19-tab-bypass-in-scheme-blocked",
+      "input":    "<a href=\"java\tscript:alert(1)\">x</a>",
+      "expected": "<a>x</a>"
+    },
+    {
+      "name": "20-uppercase-tag-and-attr-normalised",
+      "input":    "<A HREF=\"https://example.com\">x</A>",
+      "expected": "<a href=\"https://example.com\">x</a>"
+    },
+    {
+      "name": "21-style-content-dropped",
+      "input":    "<style>body{display:none}</style>visible",
+      "expected": "visible"
+    },
+    {
+      "name": "22-object-content-dropped",
+      "input":    "<object data=\"x.swf\">flash</object>safe",
+      "expected": "safe"
+    },
+    {
+      "name": "23-svg-onload-dropped",
+      "input":    "<svg onload=\"alert(1)\"><circle r=\"5\"/></svg>safe",
+      "expected": "safe"
+    },
+    {
+      "name": "24-blockquote-passthrough",
+      "input":    "<blockquote>quoted text</blockquote>",
+      "expected": "<blockquote>quoted text</blockquote>"
+    },
+    {
+      "name": "25-headings-passthrough",
+      "input":    "<h1>title</h1><h2>section</h2>",
+      "expected": "<h1>title</h1><h2>section</h2>"
+    },
+    {
+      "name": "26-attribute-value-with-gt-byte",
+      "input":    "<a href=\"https://example.com/?q=1>2\" title=\"a > b\">x</a>",
+      "expected": "<a href=\"https://example.com/?q=1&gt;2\" title=\"a &gt; b\">x</a>"
+    },
+    {
+      "name": "27-nested-script-after-text",
+      "input":    "lead<p>para</p><script>alert(1)</script>tail",
+      "expected": "lead<p>para</p>tail"
+    },
+    {
+      "name": "28-empty-input",
+      "input":    "",
+      "expected": ""
+    },
+    {
+      "name": "29-plain-text",
+      "input":    "just text, no tags",
+      "expected": "just text, no tags"
+    }
+  ]
+}
@@ -0,0 +1,32 @@
+#!/usr/bin/env bash
+# run.sh — build and execute the html_sanitizer acceptance corpus.
+#
+# Compiles tests/html_sanitizer/runner.el via the canonical native elc,
+# links against the shared C runtime, then runs the binary against
+# cases.json. Exits non-zero on any FAIL.
+
+set -euo pipefail
+cd "$(dirname "$0")"
+
+EL_HOME="${EL_HOME:-$(cd ../.. && pwd)}"
+ELC="${EL_HOME}/dist/platform/elc"
+RUNTIME_DIR="${EL_HOME}/el-compiler/runtime"
+
+if [ ! -x "${ELC}" ]; then
+    echo "elc not found at ${ELC}" >&2
+    exit 1
+fi
+
+OUT_C="$(mktemp -t html_sanitizer.XXXXXX).c"
+OUT_BIN="$(mktemp -t html_sanitizer.XXXXXX)"
+trap 'rm -f "${OUT_C}" "${OUT_BIN}"' EXIT
+
+echo "==> Compiling runner.el via ${ELC}"
+"${ELC}" runner.el > "${OUT_C}"
+
+echo "==> Linking against ${RUNTIME_DIR}/el_runtime.c"
+cc -O2 -I "${RUNTIME_DIR}" "${OUT_C}" "${RUNTIME_DIR}/el_runtime.c" \
+   -lcurl -lpthread -o "${OUT_BIN}"
+
+echo "==> Running"
+"${OUT_BIN}"
@@ -0,0 +1,59 @@
+// runner.el — acceptance tests for el_html_sanitize.
+//
+// Reads cases.json from the cwd, runs every case through
+// el_html_sanitize, prints PASS/FAIL per case and a summary,
+// and exits non-zero if any case failed.
+//
+// cases.json shape:
+//   {
+//     "allowlist": "<json string passed to el_html_sanitize>",
+//     "cases": [
+//       { "name": "...", "input": "...", "expected": "..." },
+//       ...
+//     ]
+//   }
+
+fn run_case(name: String, input: String, expected: String, allowlist: String, idx: Int) -> Int {
+    let got: String = el_html_sanitize(input, allowlist)
+    if str_eq(got, expected) {
+        println("PASS  " + name)
+        return 1
+    } else {
+        println("FAIL  " + name)
+        println("      input:    " + input)
+        println("      expected: " + expected)
+        println("      got:      " + got)
+        return 0
+    }
+}
+
+fn main() -> Int {
+    let raw: String = fs_read("cases.json")
+    if str_eq(raw, "") {
+        println("FATAL  could not read cases.json from cwd")
+        return 2
+    }
+    let allowlist: String = json_get(raw, "allowlist")
+    let cases_raw: String = json_get_raw(raw, "cases")
+    let n: Int = json_array_len(cases_raw)
+    let i: Int = 0
+    let pass: Int = 0
+    let fail: Int = 0
+    while i < n {
+        let case_obj: String = json_array_get(cases_raw, i)
+        let cname: String = json_get(case_obj, "name")
+        let cinp: String = json_get(case_obj, "input")
+        let cexp: String = json_get(case_obj, "expected")
+        let r: Int = run_case(cname, cinp, cexp, allowlist, i)
+        if r == 1 {
+            let pass = pass + 1
+        } else {
+            let fail = fail + 1
+        }
+        let i = i + 1
+    }
+    println("")
+    println(int_to_str(pass) + " passed, " + int_to_str(fail) + " failed")
+    if fail > 0 { return 1 }
+    return 0
+}