sunchao commented on code in PR #5872:
URL: https://github.com/apache/datafusion-comet/pull/5872#discussion_r4433048460
##########
native/core/src/parquet/objectstore/s3.rs:
##########
@@ -384,11 +403,42 @@ fn extract_s3_config_options(
s3_configs
}
+/// Mirrors the AWS SDK's `isVirtualHostableS3Bucket`: 3 to 63 lowercase
letters, digits and
+/// hyphens that start and end with a letter or digit. Dots only over plain
HTTP, since a dotted
+/// host falls outside S3's wildcard certificate, and then no IPv4 shape or
adjacent separators.
+fn is_virtual_hostable_bucket(bucket: &str, allow_dots: bool) -> bool {
+ let bytes = bucket.as_bytes();
+ let edge = |b: &u8| b.is_ascii_lowercase() || b.is_ascii_digit();
+ let inner = |b: &u8| edge(b) || *b == b'-' || (allow_dots && *b == b'.');
+ if !(3..=63).contains(&bytes.len())
+ || !bytes.first().is_some_and(edge)
+ || !bytes.last().is_some_and(edge)
+ || !bytes.iter().all(inner)
+ {
+ return false;
+ }
+ let ipv4_shaped = bucket.split('.').count() == 4
+ && bucket
+ .split('.')
+ .all(|label| label.bytes().all(|b| b.is_ascii_digit()));
+ let separators_touch = bytes
+ .windows(2)
+ .any(|pair| pair.iter().all(|b| *b == b'.' || *b == b'-'));
Review Comment:
For bucket `a--b`, endpoint `http://storage.example.test`, and path-style
disabled, the AWS SDK 2.31.51 endpoint provider returns
`http://a--b.storage.example.test`. This predicate also rejects adjacent
hyphens, so Comet instead requests `http://storage.example.test/a--b`. That
changes routing on a virtual-host-only endpoint. Could the separator check
allow `--` and add this case to the final-URL tests?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
