S3 client tries to connect to http service even if endpoint_url specifies full url with https

[swapnilsky]

Describe the bug

While creating client for s3, I am giving full endpoint url like https://<service-external-ip> and use_ssl as False.
according to the Documentation,
when endpoint_url specifies http scheme, botocore will ignore use_ssl.
But this isn't happening, The created client tries to connect to http://<service-external-ip> and fails.

Expected Behavior

Client should connect to endpoint_url as is when http scheme is given such as https://<service-external-ip>

Current Behavior

Client tries to connect to http://<service-external-ip>

Reproduction Steps

`
import boto3

use_ssl = False
endpoint_url = "https://" // this is dynamic, we load this from ENV VARIABLE
s3 = boto3.client("s3", aws_access_key_id="aws-access-key", aws_secret_access_key="aws-secret-access-key", use_ssl=use_ssl, endpoint_url=endpoint_url)
lock_config = s3.get_object_lock_configuration(Bucket="bucket_name")
`

Possible Solution

No response

Additional Information/Context

No response

SDK version used

1.17.99

Environment details (OS name and version, etc.)

ubuntu 20.04, python 3, boto3 1.17.99

[tim-finnigan]

Thanks for reaching out. I can reproduce this issue, for example:

import boto3
from botocore.config import Config

s3 = boto3.client("s3", endpoint_url='https://test.com', use_ssl=False,
                  config=Config(connect_timeout=1, retries={'max_attempts': 1}))
lock_config = s3.get_object_lock_configuration(Bucket='bucket_name')

Results in trying to connect to http://test.com/bucket_name?object-lock.

Relevant code can be found here. So regarding the documentation, the issue is with this wording for endpoint_url:

You can specify a complete URL (including the “http/https” scheme) to override this behavior. If this value is provided, then use_ssl is ignored.

---

I should also note that 1.17.99 is very old (the latest version of Boto3 is 1.35.12), so I recommend updating to a more recent version.