Add pytest test suite

174 tests covering URL normalization, FTS5 query sanitization, SSRF/CSRF
guards, sharing-mode logic, DB schema and upsert paths, handler
end-to-end flows, and gateway body-size / mesh-whitelist guards. Each
recent bug-fix commit (6ffd38d, 1bc695f, 8dffd8c) has an explicit
regression test in test_regressions.py. One xfail documents a minor
latent bug in clean_url where port 80 is not stripped from upgraded
https URLs.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

conftest.py

@@ -0,0 +1,128 @@
+"""Shared pytest fixtures for TinyWeb tests.
+
+Three fixtures cover most tests: `temp_db` swaps the SQLite path to a
+per-test tempfile, `seeded_db` layers sample rows on top, and `csrf_session`
+primes the thread-local CSRF token that handlers read.
+"""
+import socket
+import sys
+from pathlib import Path
+
+import pytest
+
+sys.path.insert(0, str(Path(__file__).parent))
+
+import db as db_module
+import handlers as handlers_module
+
+
+@pytest.fixture
+def temp_db(tmp_path, monkeypatch):
+    """Isolated SQLite DB per test.
+
+    Swaps `db.DATABASE` and `db.DATA_DIR` to a tempdir, clears the connection
+    pool before and after so state doesn't leak across tests, and calls
+    `init_db()` so every schema object exists.
+    """
+    data_dir = tmp_path / "tinyweb"
+    data_dir.mkdir()
+    db_path = data_dir / "index.db"
+
+    monkeypatch.setattr(db_module, "DATA_DIR", str(data_dir))
+    monkeypatch.setattr(db_module, "DATABASE", str(db_path))
+
+    with db_module._pool_lock:
+        for conn in db_module._pool:
+            try:
+                conn.close()
+            except Exception:
+                pass
+        db_module._pool.clear()
+
+    db_module.init_db()
+    yield db_path
+
+    with db_module._pool_lock:
+        for conn in db_module._pool:
+            try:
+                conn.close()
+            except Exception:
+                pass
+        db_module._pool.clear()
+
+
+@pytest.fixture
+def seeded_db(temp_db):
+    """A temp DB with a small, realistic set of pages/tags/links."""
+    db = db_module.get_db()
+    try:
+        rows = [
+            ("https://example.com/rust-intro", "Rust Intro", "A gentle introduction to rust borrow checker.", "notes on ownership"),
+            ("https://example.com/python-tips", "Python Tips", "Daily python tricks for readable code.", ""),
+            ("https://example.com/ocaml-why", "Why OCaml", "Type systems and inference in ocaml.", "private thoughts"),
+            ("https://news.example.org/mesh", "Mesh Networking", "Reticulum and LoRa for decentralized networks.", ""),
+        ]
+        for url, title, body, note in rows:
+            db.execute(
+                "INSERT INTO pages (url, title, body, note, last_modified) "
+                "VALUES (?, ?, ?, ?, '2026-04-01T00:00:00')",
+                (url, title, body, note),
+            )
+        db.commit()
+        page_ids = {
+            row["url"]: row["id"]
+            for row in db.execute("SELECT id, url FROM pages").fetchall()
+        }
+        tag_rows = [
+            (page_ids["https://example.com/rust-intro"], ["rust", "public"]),
+            (page_ids["https://example.com/python-tips"], ["python"]),
+            (page_ids["https://example.com/ocaml-why"], ["ocaml", "private"]),
+            (page_ids["https://news.example.org/mesh"], ["mesh", "public"]),
+        ]
+        for pid, tags in tag_rows:
+            for name in tags:
+                db.execute("INSERT OR IGNORE INTO tags (name) VALUES (?)", (name,))
+                tid = db.execute("SELECT id FROM tags WHERE name = ?", (name,)).fetchone()[0]
+                db.execute(
+                    "INSERT OR IGNORE INTO page_tags (page_id, tag_id) VALUES (?, ?)",
+                    (pid, tid),
+                )
+        db.execute(
+            "INSERT INTO links (page_id, url, label) VALUES (?, ?, ?)",
+            (page_ids["https://example.com/rust-intro"], "https://example.com/rust-advanced", "advanced rust guide"),
+        )
+        db.commit()
+    finally:
+        db_module.return_db(db)
+    return temp_db
+
+
+@pytest.fixture
+def csrf_session(monkeypatch):
+    """Prime the CSRF thread-local so handler code that calls _get_csrf_token works."""
+    token = "test-csrf-token"
+    handlers_module._request_local.csrf_token = token
+    yield token
+    if hasattr(handlers_module._request_local, "csrf_token"):
+        del handlers_module._request_local.csrf_token
+
+
+def patch_dns_fail(monkeypatch):
+    """Make every socket.getaddrinfo call raise gaierror for the rest of this test."""
+    def boom(*args, **kwargs):
+        raise socket.gaierror("test: DNS disabled")
+    monkeypatch.setattr(socket, "getaddrinfo", boom)
+
+
+def patch_dns_ok(monkeypatch, address="93.184.216.34"):
+    """Make every getaddrinfo return a single public IP for the rest of this test."""
+    def ok(host, port, *args, **kwargs):
+        return [(socket.AF_INET, socket.SOCK_STREAM, 0, "", (address, port or 80))]
+    monkeypatch.setattr(socket, "getaddrinfo", ok)
+
+
+def patch_dns_private(monkeypatch, address="127.0.0.1"):
+    """Make every getaddrinfo return a private/blocked IP for the rest of this test."""
+    def private(host, port, *args, **kwargs):
+        return [(socket.AF_INET, socket.SOCK_STREAM, 0, "", (address, port or 80))]
+    monkeypatch.setattr(socket, "getaddrinfo", private)