v5.42.1

README.rst

@@ -220,6 +220,14 @@ in addition to its normal output.  If these warnings can interfere with your
 scripts or you otherwise want them disabled, simply add the ``--suppress-warnings``
 flag to prevent them from being emitted.
 
+Suppressing Warnings
+""""""""""""""""""""
+
+Sometimes the API responds with a error that can be ignored. For example a timeout
+or nginx response that can't be parsed correctly, by default the CLI will retry
+calls on these errors we've identified. If you'd like to disable this behavior for
+any reason use the ``--no-retry`` flag.
+
 Shell Completion
 """"""""""""""""
 

linodecli/__init__.py

@@ -83,6 +83,8 @@ def main():  # pylint: disable=too-many-branches,too-many-statements
         cli.output_handler.columns = parsed.format
 
     cli.defaults = not parsed.no_defaults
+    cli.retry_count = 0
+    cli.no_retry = parsed.no_retry
     cli.suppress_warnings = parsed.suppress_warnings
 
     if parsed.all_columns or parsed.all:

linodecli/api_request.py

@@ -5,6 +5,7 @@
 import itertools
 import json
 import sys
+import time
 from sys import version_info
 from typing import Iterable, List, Optional
 
@@ -92,6 +93,11 @@ def do_request(
     if ctx.debug_request:
         _print_response_debug_info(result)
 
+    while _check_retry(result) and not ctx.no_retry and ctx.retry_count < 3:
+        time.sleep(_get_retry_after(result.headers))
+        ctx.retry_count += 1
+        result = method(url, headers=headers, data=body, verify=API_CA_PATH)
+
     _attempt_warn_old_version(ctx, result)
 
     if not 199 < result.status_code < 399 and not skip_error_handling:
@@ -361,3 +367,24 @@ def _handle_error(ctx, response):
             columns=["field", "reason"],
         )
     sys.exit(1)
+
+
+def _check_retry(response):
+    """
+    Check for valid retry scenario, returns true if retry is valid
+    """
+    if response.status_code in (408, 429):
+        # request timed out or rate limit exceeded
+        return True
+
+    return (
+        response.headers
+        and response.status_code == 400
+        and response.headers.get("Server") == "nginx"
+        and response.headers.get("Content-Type") == "text/html"
+    )
+
+
+def _get_retry_after(headers):
+    retry_str = headers.get("Retry-After", "")
+    return int(retry_str) if retry_str else 0

linodecli/arg_helpers.py

@@ -16,7 +16,7 @@
 from linodecli import plugins
 
 from .completion import bake_completions
-from .helpers import register_args_shared
+from .helpers import pagination_args_shared, register_args_shared
 
 
 def register_args(parser):
@@ -76,21 +76,6 @@ def register_args(parser):
         action="store_true",
         help="If set, does not display headers in output.",
     )
-    parser.add_argument(
-        "--page",
-        metavar="PAGE",
-        default=1,
-        type=int,
-        help="For listing actions, specifies the page to request",
-    )
-    parser.add_argument(
-        "--page-size",
-        metavar="PAGESIZE",
-        default=100,
-        type=int,
-        help="For listing actions, specifies the number of items per page, "
-        "accepts any value between 25 and 500",
-    )
     parser.add_argument(
         "--all",
         action="store_true",
@@ -126,6 +111,11 @@ def register_args(parser):
         default=False,
         help="Prevent the truncation of long values in command outputs.",
     )
+    parser.add_argument(
+        "--no-retry",
+        action="store_true",
+        help="Skip retrying on common errors like timeouts.",
+    )
     parser.add_argument(
         "--column-width",
         type=int,
@@ -143,6 +133,7 @@ def register_args(parser):
         "--debug", action="store_true", help="Enable verbose HTTP debug output."
     )
 
+    pagination_args_shared(parser)
     register_args_shared(parser)
 
     return parser

linodecli/baked/request.py

@@ -99,7 +99,7 @@ def _parse_request_model(schema, prefix=None, list_of_objects=False):
 
     if schema.properties is not None:
         for k, v in schema.properties.items():
-            if v.type == "object":
+            if v.type == "object" and not v.readOnly and v.properties:
                 # nested objects receive a prefix and are otherwise parsed normally
                 pref = prefix + "." + k if prefix else k
                 args += _parse_request_model(v, prefix=pref)

linodecli/baked/response.py

@@ -187,13 +187,14 @@ def __init__(self, response):
             )
         else:
             self.attrs = _parse_response_model(response.schema)
-        self.rows = response.schema.extensions.get("linode-cli-rows")
+        self.rows = response.extensions.get("linode-cli-rows")
         self.nested_list = response.extensions.get("linode-cli-nested-list")
 
     def fix_json(self, json):
         """
         Formats JSON from the API into a list of rows
         """
+
         if self.rows:
             return self._fix_json_rows(json)
         if self.nested_list:

linodecli/helpers.py

@@ -63,6 +63,33 @@ def filter_markdown_links(text):
     return result
 
 
+def pagination_args_shared(parser: ArgumentParser):
+    """
+    Add pagination related arguments to the given
+    ArgumentParser that may be shared across the CLI and plugins.
+    """
+    parser.add_argument(
+        "--page",
+        metavar="PAGE",
+        default=1,
+        type=int,
+        help="For listing actions, specifies the page to request",
+    )
+    parser.add_argument(
+        "--page-size",
+        metavar="PAGESIZE",
+        default=100,
+        type=int,
+        help="For listing actions, specifies the number of items per page, "
+        "accepts any value between 25 and 500",
+    )
+    parser.add_argument(
+        "--all-rows",
+        action="store_true",
+        help="Output all possible rows in the results with pagination",
+    )
+
+
 def register_args_shared(parser: ArgumentParser):
     """
     Adds certain arguments to the given ArgumentParser that may be shared across
@@ -87,12 +114,6 @@ def register_args_shared(parser: ArgumentParser):
         "This is useful for scripting the CLI's behavior.",
     )
 
-    parser.add_argument(
-        "--all-rows",
-        action="store_true",
-        help="Output all possible rows in the results with pagination",
-    )
-
     return parser
 
 

linodecli/plugins/obj/__init__.py

@@ -14,16 +14,15 @@
 from contextlib import suppress
 from datetime import datetime
 from math import ceil
-from pathlib import Path
-from typing import List
+from typing import Iterable, List
 
 from rich import print as rprint
 from rich.table import Table
 
 from linodecli.cli import CLI
 from linodecli.configuration import _do_get_request
 from linodecli.configuration.helpers import _default_thing_input
-from linodecli.helpers import expand_globs
+from linodecli.helpers import expand_globs, pagination_args_shared
 from linodecli.plugins import PluginContext, inherit_plugin_args
 from linodecli.plugins.obj.buckets import create_bucket, delete_bucket
 from linodecli.plugins.obj.config import (
@@ -70,14 +69,36 @@
 except ImportError:
     HAS_BOTO = False
 
+TRUNCATED_MSG = (
+    "Notice: Not all results were shown. If your would "
+    "like to get more results, you can add the '--all-row' "
+    "flag to the command or use the built-in pagination flags."
+)
+
+INVALID_PAGE_MSG = "No result to show in this page."
+
+
+def flip_to_page(iterable: Iterable, page: int = 1):
+    """Given a iterable object and return a specific iteration (page)"""
+    iterable = iter(iterable)
+    for _ in range(page - 1):
+        try:
+            next(iterable)
+        except StopIteration:
+            print(INVALID_PAGE_MSG)
+            sys.exit(2)
+
+    return next(iterable)
+
 
 def list_objects_or_buckets(
     get_client, args, **kwargs
-):  # pylint: disable=too-many-locals,unused-argument
+):  # pylint: disable=too-many-locals,unused-argument,too-many-branches
     """
     Lists buckets or objects
     """
     parser = inherit_plugin_args(ArgumentParser(PLUGIN_BASE + " ls"))
+    pagination_args_shared(parser)
 
     parser.add_argument(
         "bucket",
@@ -106,16 +127,30 @@ def list_objects_or_buckets(
             prefix = ""
 
         data = []
+        objects = []
+        sub_directories = []
+        pages = client.get_paginator("list_objects_v2").paginate(
+            Prefix=prefix,
+            Bucket=bucket_name,
+            Delimiter="/",
+            PaginationConfig={"PageSize": parsed.page_size},
+        )
         try:
-            response = client.list_objects_v2(
-                Prefix=prefix, Bucket=bucket_name, Delimiter="/"
-            )
+            if parsed.all_rows:
+                results = pages
+            else:
+                page = flip_to_page(pages, parsed.page)
+                if page.get("IsTruncated", False):
+                    print(TRUNCATED_MSG)
+
+                results = [page]
         except client.exceptions.NoSuchBucket:
             print("No bucket named " + bucket_name)
             sys.exit(2)
 
-        objects = response.get("Contents", [])
-        sub_directories = response.get("CommonPrefixes", [])
+        for item in results:
+            objects.extend(item.get("Contents", []))
+            sub_directories.extend(item.get("CommonPrefixes", []))
 
         for d in sub_directories:
             data.append((" " * 16, "DIR", d.get("Prefix")))
@@ -329,17 +364,31 @@ def list_all_objects(
     """
     # this is for printing help when --help is in the args
     parser = inherit_plugin_args(ArgumentParser(PLUGIN_BASE + " la"))
+    pagination_args_shared(parser)
 
-    parser.parse_args(args)
+    parsed = parser.parse_args(args)
 
     client = get_client()
 
-    # all buckets
     buckets = [b["Name"] for b in client.list_buckets().get("Buckets", [])]
 
     for b in buckets:
         print()
-        objects = client.list_objects_v2(Bucket=b).get("Contents", [])
+        objects = []
+        pages = client.get_paginator("list_objects_v2").paginate(
+            Bucket=b, PaginationConfig={"PageSize": parsed.page_size}
+        )
+        if parsed.all_rows:
+            results = pages
+        else:
+            page = flip_to_page(pages, parsed.page)
+            if page.get("IsTruncated", False):
+                print(TRUNCATED_MSG)
+
+            results = [page]
+
+        for page in results:
+            objects.extend(page.get("Contents", []))
 
         for obj in objects:
             size = obj.get("Size", 0)

tests/integration/helpers.py

@@ -159,3 +159,7 @@ def exec_failing_test_command(args: List[str]):
     )
     assert process.returncode == 1
     return process
+
+
+def count_lines(text: str):
+    return len(list(filter(len, text.split("\n"))))