fix(fetch-page): pass --compressed only when curl supports it (#759)

Fixes #705. Probes curl --version once per process and leaves --compressed out when the Features line lists no decoder (libz, brotli, zstd), so the System32 curl on Windows no longer rejects every fetch. Same design as #713 by @DevarshP610, which was opened first.

tests/test_fetch_page.py

@@ -349,6 +349,16 @@ class Fetching(FetchPageFixture):
         self.assertEqual(self.out.read_text(encoding="utf-8"), PAGE)
         self.assertIn("Visible title", out)
 
+    def test_fetch_succeeds_when_curl_lacks_compression(self):
+        orig = fetch_page._curl_supports_compressed
+        fetch_page._curl_supports_compressed = lambda: False
+        self.addCleanup(setattr, fetch_page, "_curl_supports_compressed", orig)
+        code, out, err = self.run_main(f"{self.base}/ok.html", "--out", str(self.out), "--text")
+        self.assertEqual(code, 0, err)
+        self.assertIn("HTTP 200 size=", out)
+        self.assertEqual(self.out.read_text(encoding="utf-8"), PAGE)
+        self.assertIn("Visible title", out)
+
 
 class CurlArgvShape(unittest.TestCase):
     """The grant is `Bash(python tools/fetch_page.py:*)`, so what curl is handed
@@ -374,6 +384,38 @@ class CurlArgvShape(unittest.TestCase):
         import robots_check
         self.assertEqual(fetch_page.BROWSER, robots_check.BROWSER)
 
+    def test_curl_compressed_flag_requires_decoder_support(self):
+        orig = fetch_page._curl_supports_compressed
+        try:
+            fetch_page._curl_supports_compressed = lambda: False
+            argv_no_decoder = fetch_page.build_curl_argv("https://example.com", Path("out.html"), Path("headers.txt"), 45)
+            self.assertNotIn("--compressed", argv_no_decoder)
+
+            fetch_page._curl_supports_compressed = lambda: True
+            argv_with_decoder = fetch_page.build_curl_argv("https://example.com", Path("out.html"), Path("headers.txt"), 45)
+            self.assertIn("--compressed", argv_with_decoder)
+        finally:
+            fetch_page._curl_supports_compressed = orig
+
+    def test_curl_features_line_controls_compressed_support(self):
+        from unittest.mock import Mock, patch
+        cases = {
+            "Features: AsynchDNS HTTPS-proxy IPv6": False,
+            "Features: AsynchDNS HTTPS-proxy IPv6 libz brotli zstd": True,
+            "Features: libz": True,
+            "Features: brotli": True,
+            "Features: zstd": True,
+        }
+        for features, expected in cases.items():
+            with self.subTest(features=features):
+                with patch.object(fetch_page.subprocess, "run", return_value=Mock(returncode=0, stdout=features)):
+                    fetch_page._curl_supports_compressed.cache_clear()
+                    self.assertEqual(fetch_page._curl_supports_compressed(), expected)
+        with patch.object(fetch_page.subprocess, "run", side_effect=OSError("not found")):
+            fetch_page._curl_supports_compressed.cache_clear()
+            self.assertFalse(fetch_page._curl_supports_compressed())
+        fetch_page._curl_supports_compressed.cache_clear()
+
 
 if __name__ == "__main__":
     unittest.main()

tools/fetch_page.py

@@ -39,6 +39,7 @@ Exit codes: 0 fetched; 1 refused by robots policy (go to escalation step 4);
 """
 
 import argparse
+from functools import lru_cache
 import gzip
 import hashlib
 import html
@@ -68,14 +69,31 @@ MAX_REDIRECTS = 5
 MAX_BYTES = 10 * 1024 * 1024
 
 
+@lru_cache(maxsize=1)
+def _curl_supports_compressed() -> bool:
+    """Return whether the bare curl used for fetching can decode content."""
+    try:
+        proc = subprocess.run(
+            ["curl", "--version"],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=5,
+        )
+    except (OSError, subprocess.SubprocessError):
+        return False
+    return proc.returncode == 0 and bool(
+        re.search(r"^Features:.*\b(?:libz|brotli|zstd)\b", proc.stdout, re.IGNORECASE | re.MULTILINE)
+    )
+
+
 def build_curl_argv(url: str, out: Path, headers: Path, timeout: float) -> list:
     """One GET without automatic redirects, so the next hop can be gated."""
+    compressed = ["--compressed"] if _curl_supports_compressed() else []
     return [
         "curl", "-sS",
         "--max-redirs", "0",
         "--max-time", str(timeout),
         "--max-filesize", str(MAX_BYTES),
-        "--compressed",
+        *compressed,
         "-A", BROWSER,
         "-H", f"Accept: {ACCEPT}",
         "-H", f"Accept-Language: {ACCEPT_LANGUAGE}",