Merge pull request #2172 from yliaog/automated-release-of-29.0.0b1-up…

…stream-release-29.0-1703785760

Automated release of 29.0.0b1 upstream release 29.0 1703785760

.github/workflows/e2e-master.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.25.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/e2e-release-11.0.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.15.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/e2e-release-12.0.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.16.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/e2e-release-17.0.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.17.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/e2e-release-18.0.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.18.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/e2e-release-26.0.yaml

@@ -30,7 +30,7 @@ jobs:
         # as we sync with Kubernetes upstream
         config: .github/workflows/kind-configs/cluster-1.26.yaml
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

.github/workflows/test.yaml

@@ -17,7 +17,7 @@ jobs:
       with:
         submodules: true
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies

CHANGELOG.md

@@ -1,3 +1,11 @@
+# v29.0.0b1
+
+Kubernetes API Version: v1.29.0
+
+### Bug or Regression
+- Fix UTF-8 failures in Watch (#2100, @davidopic)
+- Fix upper version boundary of urllib3, since other dependencies don't support urllib3 in version 2 (#2105, @jsaalfeld)
+
 # v29.0.0a1
 
 Kubernetes API Version: v1.29.0

kubernetes/README.md

@@ -4,7 +4,7 @@ No description provided (generated by Openapi Generator https://github.com/opena
 This Python package is automatically generated by the [OpenAPI Generator](https://openapi-generator.tech) project:
 
 - API version: release-1.29
-- Package version: 29.0.0a1
+- Package version: 29.0.0b1
 - Build package: org.openapitools.codegen.languages.PythonClientCodegen
 
 ## Requirements.

kubernetes/__init__.py

@@ -14,7 +14,7 @@
 
 __project__ = 'kubernetes'
 # The version is auto-updated. Please do not edit.
-__version__ = "29.0.0a1"
+__version__ = "29.0.0b1"
 
 from . import client
 from . import config

kubernetes/base/watch/watch.py

@@ -52,20 +52,33 @@ def _find_return_type(func):
 
 
 def iter_resp_lines(resp):
-    prev = ""
-    for seg in resp.stream(amt=None, decode_content=False):
-        if isinstance(seg, bytes):
-            seg = seg.decode('utf8')
-        seg = prev + seg
-        lines = seg.split("\n")
-        if not seg.endswith("\n"):
-            prev = lines[-1]
-            lines = lines[:-1]
+    buffer = bytearray()
+    for segment in resp.stream(amt=None, decode_content=False):
+
+        # Append the segment (chunk) to the buffer
+        #
+        # Performance note: depending on contents of buffer and the type+value of segment,
+        # encoding segment into the buffer could be a wasteful step. The approach used here
+        # simplifies the logic farther down, but in the future it may be reasonable to
+        # sacrifice readability for performance.
+        if isinstance(segment, bytes):
+            buffer.extend(segment)
+        elif isinstance(segment, str):
+            buffer.extend(segment.encode("utf-8"))
         else:
-            prev = ""
-        for line in lines:
+            raise TypeError(
+                f"Received invalid segment type, {type(segment)}, from stream. Accepts only 'str' or 'bytes'.")
+
+        # Split by newline (safe for utf-8 because multi-byte sequences cannot contain the newline byte)
+        next_newline = buffer.find(b'\n')
+        while next_newline != -1:
+            # Convert bytes to a valid utf-8 string, replacing any invalid utf-8 with the '�' character
+            line = buffer[:next_newline].decode(
+                "utf-8", errors="replace")
+            buffer = buffer[next_newline+1:]
             if line:
                 yield line
+            next_newline = buffer.find(b'\n')
 
 
 class Watch(object):

kubernetes/base/watch/watch_test.py

@@ -61,13 +61,133 @@ def test_watch_with_decode(self):
             if count == 4:
                 w.stop()
 
+        # make sure that all three records were consumed by the stream
+        self.assertEqual(4, count)
+
         fake_api.get_namespaces.assert_called_once_with(
             _preload_content=False, watch=True)
         fake_resp.stream.assert_called_once_with(
             amt=None, decode_content=False)
         fake_resp.close.assert_called_once()
         fake_resp.release_conn.assert_called_once()
 
+    def test_watch_with_interspersed_newlines(self):
+        fake_resp = Mock()
+        fake_resp.close = Mock()
+        fake_resp.release_conn = Mock()
+        fake_resp.stream = Mock(
+            return_value=[
+                '\n',
+                '{"type": "ADDED", "object": {"metadata":',
+                '{"name": "test1","resourceVersion": "1"}}}\n{"type": "ADDED", ',
+                '"object": {"metadata": {"name": "test2", "resourceVersion": "2"}}}\n',
+                '\n',
+                '',
+                '{"type": "ADDED", "object": {"metadata": {"name": "test3", "resourceVersion": "3"}}}\n',
+                '\n\n\n',
+                '\n',
+            ])
+
+        fake_api = Mock()
+        fake_api.get_namespaces = Mock(return_value=fake_resp)
+        fake_api.get_namespaces.__doc__ = ':return: V1NamespaceList'
+
+        w = Watch()
+        count = 0
+
+        # Consume all test events from the mock service, stopping when no more data is available.
+        # Note that "timeout_seconds" below is not a timeout; rather, it disables retries and is
+        # the only way to do so. Without that, the stream will re-read the test data forever.
+        for e in w.stream(fake_api.get_namespaces, timeout_seconds=1):
+            count += 1
+            self.assertEqual("test%d" % count, e['object'].metadata.name)
+        self.assertEqual(3, count)
+
+    def test_watch_with_multibyte_utf8(self):
+        fake_resp = Mock()
+        fake_resp.close = Mock()
+        fake_resp.release_conn = Mock()
+        fake_resp.stream = Mock(
+            return_value=[
+                # two-byte utf-8 character
+                '{"type":"MODIFIED","object":{"data":{"utf-8":"© 1"},"metadata":{"name":"test1","resourceVersion":"1"}}}\n',
+                # same copyright character expressed as bytes
+                b'{"type":"MODIFIED","object":{"data":{"utf-8":"\xC2\xA9 2"},"metadata":{"name":"test2","resourceVersion":"2"}}}\n'
+                # same copyright character with bytes split across two stream chunks
+                b'{"type":"MODIFIED","object":{"data":{"utf-8":"\xC2',
+                b'\xA9 3"},"metadata":{"n',
+                # more chunks of the same event, sent as a mix of bytes and strings
+                'ame":"test3","resourceVersion":"3"',
+                '}}}',
+                b'\n'
+            ])
+
+        fake_api = Mock()
+        fake_api.get_configmaps = Mock(return_value=fake_resp)
+        fake_api.get_configmaps.__doc__ = ':return: V1ConfigMapList'
+
+        w = Watch()
+        count = 0
+
+        # Consume all test events from the mock service, stopping when no more data is available.
+        # Note that "timeout_seconds" below is not a timeout; rather, it disables retries and is
+        # the only way to do so. Without that, the stream will re-read the test data forever.
+        for event in w.stream(fake_api.get_configmaps, timeout_seconds=1):
+            count += 1
+            self.assertEqual("MODIFIED", event['type'])
+            self.assertEqual("test%d" % count, event['object'].metadata.name)
+            self.assertEqual("© %d" % count, event['object'].data["utf-8"])
+            self.assertEqual(
+                "%d" % count, event['object'].metadata.resource_version)
+            self.assertEqual("%d" % count, w.resource_version)
+        self.assertEqual(3, count)
+
+    def test_watch_with_invalid_utf8(self):
+        fake_resp = Mock()
+        fake_resp.close = Mock()
+        fake_resp.release_conn = Mock()
+        fake_resp.stream = Mock(
+            # test 1 uses 1 invalid utf-8 byte
+            # test 2 uses a sequence of 2 invalid utf-8 bytes
+            # test 3 uses a sequence of 3 invalid utf-8 bytes
+            return_value=[
+                # utf-8 sequence for 😄 is \xF0\x9F\x98\x84
+                # all other sequences below are invalid
+                # ref: https://www.w3.org/2001/06/utf-8-wrong/UTF-8-test.html
+                b'{"type":"MODIFIED","object":{"data":{"utf-8":"\xF0\x9F\x98\x84 1","invalid":"\x80 1"},"metadata":{"name":"test1"}}}\n',
+                b'{"type":"MODIFIED","object":{"data":{"utf-8":"\xF0\x9F\x98\x84 2","invalid":"\xC0\xAF 2"},"metadata":{"name":"test2"}}}\n',
+                # mix bytes/strings and split byte sequences across chunks
+                b'{"type":"MODIFIED","object":{"data":{"utf-8":"\xF0\x9F\x98',
+                b'\x84 ',
+                b'',
+                b'3","invalid":"\xE0\x80',
+                b'\xAF ',
+                '3"},"metadata":{"n',
+                'ame":"test3"',
+                '}}}',
+                b'\n'
+            ])
+
+        fake_api = Mock()
+        fake_api.get_configmaps = Mock(return_value=fake_resp)
+        fake_api.get_configmaps.__doc__ = ':return: V1ConfigMapList'
+
+        w = Watch()
+        count = 0
+
+        # Consume all test events from the mock service, stopping when no more data is available.
+        # Note that "timeout_seconds" below is not a timeout; rather, it disables retries and is
+        # the only way to do so. Without that, the stream will re-read the test data forever.
+        for event in w.stream(fake_api.get_configmaps, timeout_seconds=1):
+            count += 1
+            self.assertEqual("MODIFIED", event['type'])
+            self.assertEqual("test%d" % count, event['object'].metadata.name)
+            self.assertEqual("😄 %d" % count, event['object'].data["utf-8"])
+            # expect N replacement characters in test N
+            self.assertEqual("� %d".replace('�', '�'*count) %
+                             count, event['object'].data["invalid"])
+        self.assertEqual(3, count)
+
     def test_watch_for_follow(self):
         fake_resp = Mock()
         fake_resp.close = Mock()

kubernetes/client/__init__.py

@@ -14,7 +14,7 @@
 
 from __future__ import absolute_import
 
-__version__ = "29.0.0a1"
+__version__ = "29.0.0b1"
 
 # import apis into sdk package
 from kubernetes.client.api.well_known_api import WellKnownApi

kubernetes/client/api_client.py

@@ -78,7 +78,7 @@ def __init__(self, configuration=None, header_name=None, header_value=None,
             self.default_headers[header_name] = header_value
         self.cookie = cookie
         # Set default User-Agent.
-        self.user_agent = 'OpenAPI-Generator/29.0.0a1/python'
+        self.user_agent = 'OpenAPI-Generator/29.0.0b1/python'
         self.client_side_validation = configuration.client_side_validation
 
     def __enter__(self):

kubernetes/client/configuration.py

@@ -354,7 +354,7 @@ def to_debug_report(self):
                "OS: {env}\n"\
                "Python Version: {pyversion}\n"\
                "Version of the API: release-1.29\n"\
-               "SDK Package Version: 29.0.0a1".\
+               "SDK Package Version: 29.0.0b1".\
                format(env=sys.platform, pyversion=sys.version)
 
     def get_host_settings(self):

scripts/constants.py

@@ -18,13 +18,13 @@
 KUBERNETES_BRANCH = "release-1.29"
 
 # client version for packaging and releasing.
-CLIENT_VERSION = "29.0.0a1"
+CLIENT_VERSION = "29.0.0b1"
 
 # Name of the release package
 PACKAGE_NAME = "kubernetes"
 
 # Stage of development, mainly used in setup.py's classifiers.
-DEVELOPMENT_STATUS = "3 - Alpha"
+DEVELOPMENT_STATUS = "4 - Beta"
 
 
 # If called directly, return the constant value given

setup.py

@@ -16,9 +16,9 @@
 
 # Do not edit these constants. They will be updated automatically
 # by scripts/update-client.sh.
-CLIENT_VERSION = "29.0.0a1"
+CLIENT_VERSION = "29.0.0b1"
 PACKAGE_NAME = "kubernetes"
-DEVELOPMENT_STATUS = "3 - Alpha"
+DEVELOPMENT_STATUS = "4 - Beta"
 
 # To install the library, run the following
 #