Parser should not set a user agent, Reader should.

src/reader/__init__.py

@@ -93,6 +93,11 @@
 _CLI_PLUGIN_ENVVAR = 'READER_CLI_PLUGIN'
 
 
+# Constants.
+
+USER_AGENT = f'python-reader/{__version__} (+https://github.com/lemon24/reader)'
+
+
 # Prevent any logging output by default. If no handler is set,
 # the messages bubble up to the root logger and get printed on stderr.
 # https://docs.python.org/3/howto/logging.html#library-config

src/reader/_parser.py

@@ -19,7 +19,6 @@
 from typing import runtime_checkable
 from typing import TypeVar
 
-import reader
 from ._http_utils import parse_accept_header
 from ._http_utils import unparse_accept_header
 from ._requests_utils import DEFAULT_TIMEOUT
@@ -40,9 +39,6 @@
 log = logging.getLogger('reader')
 
 
-USER_AGENT = f'python-reader/{reader.__version__} (+https://github.com/lemon24/reader)'
-
-
 def default_parser(
     feed_root: str | None = None, session_timeout: TimeoutType = DEFAULT_TIMEOUT
 ) -> Parser:
@@ -126,7 +122,7 @@ def __init__(self) -> None:
         #:
         #: Plugins may add request or response hooks to this.
         #:
-        self.session_factory = SessionFactory(USER_AGENT)
+        self.session_factory = SessionFactory()
 
     def parallel(
         self,

src/reader/core.py

@@ -230,6 +230,10 @@ def make_reader(
         raise ValueError("search_enabled should be one of ('auto', True, False, None)")
 
     parser = default_parser(feed_root, session_timeout=session_timeout)
+    # circular import
+    from . import USER_AGENT
+
+    parser.session_factory.user_agent = USER_AGENT
 
     try:
         name_scheme = NameScheme.from_value(reserved_name_scheme)

tests/test_parser.py

@@ -439,12 +439,12 @@ def get(self, *args, **kwargs):
     assert 'while getting feed' in excinfo.value.message
 
 
-def test_user_agent(parse, make_http_get_headers_url, data_dir):
+def test_user_agent_default(parse, make_http_get_headers_url, data_dir):
     feed_url = make_http_get_headers_url(data_dir.join('full.atom'))
     parse(feed_url)
 
     headers = make_http_get_headers_url.request_headers
-    assert headers['User-Agent'].startswith('python-reader/')
+    assert headers['User-Agent'].startswith('python-requests/')
 
 
 def test_user_agent_none(parse, make_http_get_headers_url, data_dir):
@@ -453,7 +453,7 @@ def test_user_agent_none(parse, make_http_get_headers_url, data_dir):
     parse(feed_url)
 
     headers = make_http_get_headers_url.request_headers
-    assert 'reader' not in headers['User-Agent']
+    assert headers['User-Agent'].startswith('python-requests/')
 
 
 def test_parallel_persistent_session(parse, make_http_url, data_dir):

tests/test_reader_integration.py

@@ -5,6 +5,7 @@
 from utils import make_url_base
 from utils import utc_datetime
 
+from reader import USER_AGENT
 from reader._types import fix_datetime_tzinfo
 
 
@@ -153,7 +154,7 @@ class datetime_mock(datetime):
 def test_etag_last_modified(reader, data_dir, server):
     """Check ETag / Last-Modified are sent back, at the wire level.
 
-    Because of https://rachelbythebay.com/w/2023/01/18/http/
+    Ensures https://rachelbythebay.com/w/2023/01/18/http/ cannot happen.
 
     """
     etag = b'"12345-67890abcdef12"'
@@ -179,4 +180,9 @@ def test_etag_last_modified(reader, data_dir, server):
     assert headers[b'If-None-Match'] == etag
     assert headers[b'If-Modified-Since'] == last_modified
 
-    # TODO: since we're here, assert user agent, accept, and a-im too
+    ua = headers[b'User-Agent']
+    assert ua == USER_AGENT.encode()
+    assert ua.startswith(b'python-reader/')
+    assert ua.endswith(b' (+https://github.com/lemon24/reader)')
+
+    # TODO: since we're here, assert accept, and a-im too; subtests, ideally