Fixing sentinel command response

[mahigupta]

Pull Request check-list

• Do tests and lints pass with this change?
• Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)?
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?
• Is there an example added to the examples folder (if applicable)?
• Was the change added to CHANGES file?

Description of change

For #3139

• Made change to execute_command function to parse sentinel command response correctly and don't return by default True

[mahigupta]

@dvora-h Can you please take a look.

[mahigupta]

@vladvildanov Can you please take a look ?

[vladvildanov]

@mahigupta Thanks for your input! I'll also include @gerzse into this

[mahigupta]

Hello,
@vladvildanov @gerzse , Can anyone look in to this PR ?

[vladvildanov]

@mahigupta Looks fine 👌 Working on fixing CI, docker-compose was removed from default commands in Ubuntu 22.04 LTS runner

[mahigupta]

@vladvildanov @gerzse Ci fixed now to go ahead with this PR ?

[petyaslavova]

This change will still need a few new unit tests to be added.

[Copilot]

Pull Request Overview

This PR fixes the behavior of the sentinel command response in Redis by updating the execute_command function to correctly parse responses and returning either a boolean or the original response based on the provided options.

• Updates in both synchronous and asynchronous sentinel modules to pop "once" and "return_responses" from kwargs and return the processed response.
• Updates to the Redis commands for sentinel to pass the new execution options.
• An update to the CHANGES file documenting the fix.

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

File	Description
redis/sentinel.py	Updated execute_command to handle return_responses and once logic.
redis/commands/sentinel.py	Modified sentinel command method calls to include new options.
redis/asyncio/sentinel.py	Asynchronous execute_command updated similar to its sync counterpart.
CHANGES	Documented the fix for the sentinel command response behavior.

[Copilot]

Pull Request Overview

This PR fixes how sentinel commands parse and return responses by introducing a return_responses flag and removing the default True return behavior. It also updates several sentinel command wrappers to leverage this new flag and expands test coverage with real-sentinel integration tests.

• Enhance execute_command (sync/async) to pop once/return_responses flags and return a list or boolean accordingly
• Update sentinel command methods to use once=True and/or return_responses=True where appropriate
• Add deployed_sentinel fixtures and real-instance tests in both sync and asyncio suites

Reviewed Changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/test_sentinel.py	Added deployed_sentinel fixture and real-sentinel integration tests
tests/test_asyncio/test_sentinel.py	Mirror sync changes for asyncio tests and real-instance coverage
redis/sentinel.py	Refactored execute_command to support return_responses flag
redis/asyncio/sentinel.py	Refactored async execute_command similarly
redis/commands/sentinel.py	Updated sentinel_* wrappers to pass once/return_responses options

Comments suppressed due to low confidence (2)

redis/commands/sentinel.py:15

• The method now returns a list of tuples (due to return_responses=True) rather than a single tuple; please update this docstring to reflect the new return type.

        """Returns a (host, port) pair for the given ``service_name``"""

redis/commands/sentinel.py:23

• This change makes sentinel_master return a list of dicts instead of a single dict, which is a breaking API change; consider documenting this behavior clearly or offering a non-breaking alternative.

def sentinel_master(self, service_name):

[Copilot]

[nitpick] Consider replacing reduce(lambda x, y: x and y, responses) with the built-in all(responses) for improved readability and to avoid the extra import.

Suggested change

	return bool(reduce(lambda x, y: x and y, responses))
	return all(responses)

[Copilot]

[nitpick] Similar to the sync version, you can use all(responses) instead of reduce to simplify the boolean aggregation.

Suggested change

	return bool(reduce(lambda x, y: x and y, responses))
	return all(responses)