fix: preserve exception subtype in request(), render params/timeout in as_curl()

request() re-raised every aiohttp.ClientError subclass as the bare base class,
losing ClientConnectorError/ClientProxyConnectionError/ClientResponseError
(.status/.headers)/TooManyRedirects and their attributes; direct callers
branching by type never matched. Now the original exception is bare-raised,
preserving subtype, attributes, and __cause__. request_with_retries (which
catches the base ClientError) is unaffected.

as_curl() silently dropped params and timeout, so a debug=True replay of a
params-driven request hit a different URL than the one actually sent. params
are now merged into the URL via yarl before quoting, and timeout renders as
--max-time.

Bumps 0.1.7 -> 0.1.8.

Signed-off-by: disqualifier <dev@disqualifier.me>

.gitignore

@@ -1,5 +1,5 @@
 # claude
-CLAUDE.md
+.claude/
 
 # python
 __pycache__/

README.md

@@ -11,17 +11,19 @@ and swap the HTTP client while inheriting everything else.
 `requirements.txt`:
 
 ```
-aioweb @ git+ssh://git@git.rethinkstudios.io/rethink-public/aioweb.git@v0.1.0
+aioweb @ git+ssh://git@git.rethinkstudios.io/rethink-public/aioweb.git@v0.1.8
 ```
 
 Direct:
 
 ```bash
-pip install "aioweb @ git+ssh://git@git.rethinkstudios.io/rethink-public/aioweb.git@v0.1.0"
+pip install "aioweb @ git+ssh://git@git.rethinkstudios.io/rethink-public/aioweb.git@v0.1.8"
 ```
 
 Requires `aiohttp` and `yarl` (pulled transitively).
 
+Drop the `@v0.1.8` suffix from the line above to install the latest unpinned.
+
 ## Usage
 
 ```python
@@ -63,6 +65,15 @@ resp = await s.request_with_retries(
 
 Returns a `FailureResponse` (falsy) if every attempt fails.
 
+`request()` (the non-retrying call) raises on failure: a total timeout raises
+`aiohttp.ServerTimeoutError`, and any other network/protocol failure raises its real
+`aiohttp.ClientError` subtype as-is (`ClientConnectorError`, `ClientProxyConnectionError`,
+`ClientResponseError` with `.status`/`.headers`, `TooManyRedirects`, ...) — it is not
+flattened into the base `ClientError`, so direct callers can branch by type or read
+subtype attributes. `request_with_retries` catches the base `aiohttp.ClientError` (and
+`asyncio.TimeoutError`) across all attempts and returns a falsy `FailureResponse` instead
+of raising.
+
 ## Header overwrites & ephemeral headers
 
 ```python
@@ -86,7 +97,10 @@ s.overwrite_domain("internal.local", "127.0.0.1")   # host-substring rewrite
 print(s.preview("POST", url, json={"a": 1}).as_curl())   # equivalent cURL command
 ```
 
-Pass `debug=True` to `request_with_retries` to log the cURL preview and request flow.
+`as_curl()` renders `params` (merged into the url's query string) and `timeout` (as
+`--max-time`) as well as headers/body/proxy, so the emitted command is faithful to what
+`request()` actually sends. Pass `debug=True` to `request_with_retries` to log the cURL
+preview and request flow.
 
 ## Custom backends
 
@@ -128,6 +142,63 @@ Two changes can't be shimmed without re-introducing the bugs they fix:
   `await s.close()`; a leaked session emits a `ResourceWarning`. The old finalizer-based
   auto-close was unsafe and was removed.
 
+## Changelog
+
+### v0.1.8
+
+- **`request()` no longer flattens `aiohttp.ClientError` subtypes.** Every failure
+  (connect errors, proxy errors, `raise_for_status()`-style response errors,
+  redirect limits, ...) was re-raised as a bare `aiohttp.ClientError`, losing the
+  real subtype and its attributes (`.os_error`, `.status`, `.headers`, ...) — a
+  direct caller doing `except ClientProxyConnectionError:` or `if e.status == 401`
+  would silently never match. Now the original exception is re-raised as-is (its
+  subtype, attributes, and `__cause__` all preserved). `request_with_retries`
+  still catches the base `aiohttp.ClientError` across attempts, so its behavior
+  (and its `FailureResponse` return on exhaustion) is unchanged.
+- **`as_curl()` now renders `params` and `timeout`.** Previously a preview built
+  with `params=` silently omitted the query string (and a `timeout=` omitted
+  `--max-time`), so a `debug=True` cURL replay of a params-driven request hit a
+  different URL than the one actually sent. `params` are now merged into the url's
+  query string (via `yarl`) and `timeout` is emitted as `--max-time`.
+
+### v0.1.7
+
+- **`get_cookies()` now returns real cookies.** Previously called `filter_cookies()`
+  with no URL, which only ever returns domain-less shared cookies — every normal
+  domain-bound cookie (including ones set by a real `Set-Cookie` response) was
+  silently omitted. Now iterates the jar directly.
+- **`set_cookie()` no longer leaks a shared cookie to every host.** A bare hostname
+  (`domain="example.com"`) built a schemeless URL, which aiohttp's jar treats as a
+  domain-less "shared" cookie sent with every request the session makes, including
+  unrelated hosts. A scheme is now added when missing so the cookie is scoped to
+  that host.
+- **`set_cookie()` now honors `path`** (previously ignored — the cookie always
+  landed at `path="/"`). `domain=None` is unchanged in meaning but now stores a
+  truly shared cookie (sent to every host) instead of one silently bound to
+  `localhost` only, which made `set_cookie(name, value)` (no domain) a silent
+  no-op for any real request.
+- **The backend session is built lazily**, not in `__init__`. Under aiohttp 3.14,
+  constructing `aiohttp.ClientSession` requires a running event loop; eager
+  construction crashed the common host pattern of attaching a session before the
+  loop starts (e.g. `bot.http = ExtendedSession(...)` in `Bot.__init__`). The
+  session (and any subclass's `_create_session` override) now builds on first
+  access instead.
+
+### v0.1.2
+
+- Pinned `commons` to v0.2.1 (retry `attempts` floor fix).
+
+### v0.1.1
+
+- **JSON list bodies** now route to `json=` (were wrongly form-encoded via `data=` —
+  only dicts went to `json=` before).
+- **Exhausted retries return the real last response.** When every attempt hit a
+  retryable status (429/5xx), the loop discarded it and returned a synthetic
+  `FailureResponse` (status 0); now the real last 4xx/5xx `Response` is returned (only a
+  pure-exception failure yields `FailureResponse`).
+- Retry/backoff moved onto `commons.aretry` (shared engine); backoff schedule unchanged.
+  Adds a `commons` dependency.
+
 ## Versioning
 
-Tagged `vX.Y.Z`. Pin the tag in `requirements.txt`.
+Releases are tagged `vX.Y.Z`. The install line above pins a release; drop the `@vX.Y.Z` suffix to install the latest unpinned. Pin deliberately for reproducible installs.

pyproject.toml

@@ -4,13 +4,17 @@ build-backend = "hatchling.build"
 
 [project]
 name = "aioweb"
-version = "0.1.0"
+version = "0.1.8"
 description = "Async HTTP session wrapper over aiohttp — proxies, header overwrites, retries, previews. Config-free, installable."
 requires-python = ">=3.10"
 dependencies = [
     "aiohttp>=3.9",
     "yarl>=1.9",
+    "commons @ git+ssh://git@git.rethinkstudios.io/rethink-public/commons.git@v0.2.1",
 ]
 
+[tool.hatch.metadata]
+allow-direct-references = true
+
 [tool.hatch.build.targets.wheel]
 packages = ["src/aioweb"]

src/aioweb/preview.py

@@ -3,6 +3,9 @@ request preview for aioweb — format or export a request without sending it
 """
 
 import json as _json
+import shlex
+
+from yarl import URL
 
 
 class RequestPreview:
@@ -25,15 +28,30 @@ class RequestPreview:
         return "\n".join(f"{key}: {value}" for key, value in self.details.items())
 
     def as_curl(self):
-        """equivalent cURL command for the request"""
+        """equivalent cURL command for the request
-        parts = [f"curl -X {self.details['method']}"]
+
+        every interpolated value is shell-quoted with shlex.quote, so headers,
+        body, url, or proxy containing quotes/spaces/metacharacters produce a
+        valid, non-injectable command rather than a broken or unsafe one. `params`
+        is merged into the url's query string (via yarl) and `timeout` is rendered
+        as `--max-time`, so the emitted command is faithful to what request()
+        actually sends rather than dropping either silently.
+        """
+        parts = [f"curl -X {shlex.quote(self.details['method'])}"]
         for header, value in (self.details["headers"] or {}).items():
-            parts.append(f"-H '{header}: {value}'")
+            parts.append(f"-H {shlex.quote(f'{header}: {value}')}")
-        if self.details["data"]:
+        if self.details["data"] is not None:
-            parts.append(f"--data '{self.details['data']}'")
+            parts.append(f"--data {shlex.quote(str(self.details['data']))}")
-        elif self.details["json"]:
+        elif self.details["json"] is not None:
-            parts.append(f"--data '{_json.dumps(self.details['json'])}'")
+            # is-not-None, not truthiness: an empty-but-valid body ({} / []) must still
-        parts.append(f"'{self.details['url']}'")
+            # render rather than being dropped as falsy
+            parts.append(f"--data {shlex.quote(_json.dumps(self.details['json']))}")
+        url = self.details["url"]
+        if self.details["params"]:
+            url = str(URL(url).update_query(self.details["params"]))
+        parts.append(shlex.quote(str(url)))
         if self.details["proxy"]:
-            parts.append(f"--proxy '{self.details['proxy']}'")
+            parts.append(f"--proxy {shlex.quote(str(self.details['proxy']))}")
+        if self.details["timeout"] is not None:
+            parts.append(f"--max-time {shlex.quote(str(self.details['timeout']))}")
         return " \\\n  ".join(parts)

src/aioweb/responses.py

@@ -92,7 +92,9 @@ class Response:
         """parsed JSON content, or None if not valid JSON"""
         try:
             return _json.loads(self.text())
-        except _json.JSONDecodeError:
+        except (_json.JSONDecodeError, UnicodeDecodeError):
+            # text() decodes the body and can raise UnicodeDecodeError on a non-UTF-8
+            # payload — that's a "not valid JSON" outcome, not an error to propagate
             return None
 
     def raise_for_status(self):

src/aioweb/session.py

@@ -12,7 +12,10 @@ subclass and override just that one method, inheriting everything else.
         if resp:                       # FailureResponse is falsy
             data = resp.json()
 
-config-free: proxies/headers/timeouts are passed at construction or per call.
+config-free: proxies/headers/timeouts are passed at construction or per call. the
+backend HTTP session is built lazily on first use (request/cookie access/close), not
+in __init__, so construction is safe before an event loop is running (e.g. attaching
+to a host object at process startup).
 sessions must be closed explicitly (async with, or await s.close()); there is no
 __del__ auto-close (that pattern is unsafe for async resources).
 """
@@ -20,13 +23,40 @@ __del__ auto-close (that pattern is unsafe for async resources).
 import asyncio
 import logging
 import warnings
+from http.cookies import SimpleCookie
 
 import aiohttp
 from yarl import URL
+from commons import aretry
 
 from .preview import RequestPreview
 from .responses import Response, FailureResponse
 
+
+def _route_body(data):
+    """split a body into (data=, json=) kwargs
+
+    dict OR list bodies are valid JSON and route to json=; everything else
+    (str/bytes/form) routes to data=. previously only dicts went to json=, so a
+    JSON list was wrongly form-encoded.
+    """
+    if isinstance(data, (dict, list)):
+        return None, data
+    return data, None
+
+
+class _RetryStatus(Exception):
+    """internal signal: a retryable HTTP status; carries the real Response
+
+    raised inside an attempt so commons.aretry drives the backoff + cap; the caller
+    catches the final one to return the REAL last response, not a synthetic failure.
+    """
+
+    def __init__(self, response):
+        super().__init__(f"retryable status {response.status_code}")
+        self.response = response
+
+
 log = logging.getLogger(__name__)
 
 DEFAULT_ATTEMPTS = 3
@@ -60,7 +90,27 @@ class ExtendedSession:
         self.proxies = proxies or {}
         # track our own default headers instead of touching aiohttp privates
         self._default_headers = dict(headers or {})
-        self.session = self._create_session(self._default_headers, timeout, **kwargs)
+        self._session_timeout = timeout
+        self._session_kwargs = kwargs
+        # aiohttp.ClientSession (via _create_session) requires a running event loop
+        # (aiohttp >= 3.14 raises RuntimeError otherwise); building it here would
+        # break the common host pattern of constructing before the loop starts
+        # (e.g. bot.http = ExtendedSession(...) in Bot.__init__). build lazily on
+        # first access instead, via the `session` property / _ensure_session().
+        self._session = None
+
+    @property
+    def session(self):
+        """the backend session, built lazily on first access (needs a running loop)"""
+        self._ensure_session()
+        return self._session
+
+    def _ensure_session(self):
+        """build the backend session on first use — idempotent"""
+        if self._session is None:
+            self._session = self._create_session(
+                self._default_headers, self._session_timeout, **self._session_kwargs,
+            )
 
     def _create_session(self, headers, timeout, **kwargs):
         """create the backend HTTP session — override to use a different client
@@ -70,9 +120,15 @@ class ExtendedSession:
         proxy/retry/preview logic in this class never touches the session object
         directly (only _raw_request, the cookie methods, and close do), so those
         features work unchanged on any backend.
+
+        `headers` is the session-default header set; the default aiohttp backend
+        does NOT bake it into the ClientSession (which would copy it into an
+        immutable per-session map that update_headers/clear_headers can't touch).
+        instead `_default_headers` is our own mutable layer that request() and
+        preview() merge per call, so the mutable session-header API actually works.
+        a backend that needs the defaults baked at construction may use `headers`.
         """
         return aiohttp.ClientSession(
-            headers=headers,
             timeout=aiohttp.ClientTimeout(
                 total=timeout,
                 connect=timeout / 2,
@@ -92,10 +148,13 @@ class ExtendedSession:
     def _apply_overwrites(self, request_headers):
         """apply static overwrites and ephemeral headers to a request's headers"""
         request_headers = dict(request_headers or {})
-        for header, value in self.header_overwrites.items():
+        # snapshot the shared override dicts so a concurrent mutation (e.g. a command
+        # editing ephemerals on a shared session) can't raise "dict changed size during
+        # iteration" — the loop body is sync, but the snapshot is cheap insurance
+        for header, value in list(self.header_overwrites.items()):
             if self.inject or header in request_headers:
                 request_headers[header] = value
-        for header, value_callable in self.ephemeral_headers.items():
+        for header, value_callable in list(self.ephemeral_headers.items()):
             if self.inject or header in request_headers:
                 value = value_callable()
                 if isinstance(value, dict):
@@ -182,13 +241,36 @@ class ExtendedSession:
     #   cookies
 
     def get_cookies(self):
-        """cookies stored in the session jar"""
+        """all cookies stored in the session jar, regardless of domain binding
-        return self.session.cookie_jar.filter_cookies()
+
+        iterates the jar directly rather than filter_cookies() (which needs a url
+        and, given none, returns only domain-less shared cookies — i.e. {} for any
+        normal domain-bound cookie).
+        """
+        return {c.key: c.value for c in self.session.cookie_jar}
 
     def set_cookie(self, name, value, domain=None, path="/"):
-        """set a cookie in the session jar"""
+        """set a cookie in the session jar
-        response_url = URL(domain or "http://localhost")
+
-        self.session.cookie_jar.update_cookies({name: value}, response_url=response_url)
+        domain=None (the default) stores a truly shared cookie sent with every
+        request regardless of host — the jar's own "no response_url" behavior.
+        pass domain='example.com' (a scheme is optional and defaulted to http://)
+        to scope the cookie to one host; a bare hostname like 'example.com' is
+        normalized into a URL so the jar binds it by host instead of silently
+        storing another domain-less shared cookie (a schemeless domain has no
+        raw_host, so the jar can't tell it apart from the shared case).
+        `path` is honored via the morsel itself, since the jar only derives a path
+        from response_url when the morsel doesn't already carry one.
+        """
+        cookie = SimpleCookie()
+        cookie[name] = value
+        if domain is None:
+            self.session.cookie_jar.update_cookies(cookie)
+            return
+        if "://" not in domain:
+            domain = "http://" + domain
+        cookie[name]["path"] = path
+        self.session.cookie_jar.update_cookies(cookie, response_url=URL(domain))
 
     def clear_cookies(self):
         """clear the session cookie jar"""
@@ -208,10 +290,8 @@ class ExtendedSession:
     def preview(self, method, url, **kwargs):
         """build a RequestPreview for a request without sending it"""
         proxy = self._get_proxy(url, kwargs.pop("proxies", None))
-        if kwargs.get("headers"):
+        merged = {**self._default_headers, **(kwargs.pop("headers", None) or {})}
-            headers = self._apply_overwrites(kwargs.pop("headers"))
+        headers = self._apply_overwrites(merged)
-        else:
-            headers = dict(self.get_headers())
 
         timeout = kwargs.get("timeout")
         timeout_total = timeout if isinstance(timeout, (int, float)) else None
@@ -262,16 +342,30 @@ class ExtendedSession:
             )
 
     async def request(self, method, url, **kwargs) -> Response:
-        """make a request, applying overwrites/domain rewrites/proxy resolution"""
+        """make a request, applying overwrites/domain rewrites/proxy resolution
+
+        raises the real exception subtype on failure: a total timeout raises
+        aiohttp.ServerTimeoutError, and any other aiohttp.ClientError (e.g.
+        ClientConnectorError, ClientProxyConnectionError, ClientResponseError with
+        .status/.headers, TooManyRedirects) is re-raised as-is, not flattened into
+        the base ClientError — callers can branch by type or read subtype attributes.
+        """
         kwargs["proxy"] = self._get_proxy(url, kwargs.pop("proxies", None))
         debug = kwargs.pop("debug", False)
 
-        kwargs["headers"] = self._apply_overwrites(kwargs.get("headers"))
+        merged = {**self._default_headers, **(kwargs.get("headers") or {})}
+        kwargs["headers"] = self._apply_overwrites(merged)
         kwargs["headers"] = {str(k): str(v) for k, v in kwargs["headers"].items()}
 
         timeout = kwargs.get("timeout")
         if isinstance(timeout, (int, float)):
             kwargs["timeout"] = aiohttp.ClientTimeout(total=timeout)
+        elif timeout is None and "timeout" in kwargs:
+            # an explicit timeout=None reaches aiohttp as ClientTimeout(total=None),
+            # which DISABLES the timeout and overrides the session default; drop it so
+            # the session-level timeout applies (matters for request_with_retries, whose
+            # timeout kwarg defaults to None)
+            del kwargs["timeout"]
 
         url = self._apply_domain_overwrites(url)
         if debug:
@@ -282,8 +376,20 @@ class ExtendedSession:
             if debug and result.redirect_chain:
                 log.info("redirect chain: %s", result.redirect_chain)
             return result
+        except asyncio.TimeoutError as error:
+            # a total ClientTimeout raises a bare asyncio.TimeoutError, which is NOT an
+            # aiohttp.ClientError subclass — wrap it as ServerTimeoutError (which IS both
+            # a ClientError AND a TimeoutError) so direct callers get a typed failure and
+            # request_with_retries can still label it a timeout
+            raise aiohttp.ServerTimeoutError(f"timeout for {url}: {error}") from error
         except aiohttp.ClientError as error:
-            raise aiohttp.ClientError(f"client error for {url}: {error}") from error
+            # re-raise the ORIGINAL subtype (ClientConnectorError, ClientProxyConnectionError,
+            # ClientResponseError with .status/.headers, TooManyRedirects, ...) instead of
+            # flattening into the base class — direct callers branching by type or reading
+            # subtype attributes need the real exception; request_with_retries still catches
+            # the base aiohttp.ClientError below and is unaffected
+            log.error("client error for %s: %s", url, error)
+            raise
 
     async def request_with_retries(
         self, method, url, *, data=None, proxies=None, timeout=None, attempts=None,
@@ -295,55 +401,76 @@ class ExtendedSession:
         returns a Response on success (or non-retryable status), or a falsy
         FailureResponse if every attempt fails. backoff is exponential
         (backoff_base ** attempt).
+
+        timeout defaults to None, which falls back to the session-level timeout set
+        at construction (aioweb pops a None timeout so it does not reach aiohttp as an
+        infinite ClientTimeout); pass a number to override per call.
         """
         attempts = attempts or DEFAULT_ATTEMPTS
-        last_error = None
+        body_data, body_json = _route_body(data)
 
         if debug:
             preview = self.preview(
                 method=method, url=url, params=params,
-                data=None if isinstance(data, dict) else data,
+                data=body_data, json=body_json,
-                json=data if isinstance(data, dict) else None,
                 headers=headers, proxies=proxies, timeout=timeout,
             ).as_curl()
             log.info("[aioweb.debug]\n%s\nproxies: %s inject: %s", preview, self.proxies, self.inject)
 
-        for attempt in range(attempts):
+        async def attempt():
-            try:
             response = await self.request(
                 method=method, url=url, params=params,
-                    data=None if isinstance(data, dict) else data,
+                data=body_data, json=body_json,
-                    json=data if isinstance(data, dict) else None,
                 headers=headers, proxies=proxies, timeout=timeout, debug=debug,
             )
             if response.status_code in retry_statuses:
-                    last_error = f"retryable status {response.status_code}"
+                log.warning("retryable status %s for %s", response.status_code, url)
-                    log.warning("attempt %d: %s for %s", attempt + 1, last_error, url)
+                raise _RetryStatus(response)
-                else:
             return response
+
+        try:
+            return await aretry(
+                attempt, attempts=attempts, backoff=1.0, factor=backoff_base,
+                jitter=False, on=(Exception,),
+            )
+        except _RetryStatus as exhausted:
+            log.error("all %d attempts failed for %s (last status %s)",
+                      attempts, url, exhausted.response.status_code)
+            return exhausted.response
+        except asyncio.TimeoutError:
+            # request() wraps a total timeout as ServerTimeoutError (a ClientError AND a
+            # TimeoutError); catch the timeout case first so it's labeled a timeout rather
+            # than falling into the generic client-error branch below
+            log.error("all %d attempts timed out for %s", attempts, url)
+            return FailureResponse(reason="timeout", url=url)
         except aiohttp.ClientError as error:
-                last_error = f"client error: {error}"
+            log.error("all %d attempts failed for %s (client error: %s)", attempts, url, error)
-                log.warning("attempt %d: %s, retrying", attempt + 1, last_error)
+            return FailureResponse(reason=f"client error: {error}", url=url)
         except Exception as error:
-                last_error = f"unexpected error: {error}"
+            log.error("all %d attempts failed for %s (unexpected: %s)", attempts, url, error)
-                log.exception("attempt %d: %s, retrying", attempt + 1, last_error)
+            return FailureResponse(reason=f"unexpected error: {error}", url=url)
-
-            if attempt < attempts - 1:
-                await asyncio.sleep(backoff_base ** attempt)
-
-        log.error("all %d attempts failed for %s", attempts, url)
-        return FailureResponse(reason=last_error, url=url)
 
     # -------------------------------------------------------------------------
     #   lifecycle
 
     async def close(self):
-        """close the backend session — override if the backend's close differs"""
+        """close the backend session — override if the backend's close differs
-        await self.session.close()
+
+        a no-op if the session was never built (lazy construction means a session
+        that made no request and was never otherwise touched has nothing to close).
+        """
+        if self._session is not None:
+            await self._session.close()
 
     def _is_closed(self) -> bool:
-        """whether the backend session is closed — override for non-aiohttp backends"""
+        """whether the backend session is closed — override for non-aiohttp backends
-        return self.session.closed
+
+        an unbuilt (never-lazily-created) session counts as closed: nothing was
+        opened, so there is nothing to leak and __del__ should not warn.
+        """
+        if self._session is None:
+            return True
+        return self._session.closed
 
     async def __aenter__(self):
         return self