Support multiple EXT-X-KEY tags

[bbayles]

The HLS spec says:

Two or
   more EXT-X-KEY tags with different KEYFORMAT attributes MAY apply to
   the same Media Segment if they ultimately produce the same decryption
   key. 

This PR updates the package such that multiple keys are supported.

If others can test this, I'll release it as a new major version!

[Copilot]

Pull request overview

This PR implements support for multiple EXT-X-KEY tags per segment in accordance with the HLS specification, which allows segments to have multiple encryption keys with different KEYFORMAT attributes for Common Encryption support.

Changes:

• Modified the parser to track multiple keys per segment instead of a single key
• Updated the Segment model to store a list of keys while maintaining backward compatibility through a key property
• Added comprehensive test coverage for parsing and round-trip serialization of multi-key playlists

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/test_parser.py	Updated tests to access segment keys via keys list and added new test for multiple KEYFORMAT support
tests/test_model.py	Added tests for segment.keys list access, backward compatibility of segment.key property, and round-trip preservation
tests/playlists.py	Added test playlist with multiple EXT-X-KEY tags using different KEYFORMAT attributes
openm3u8/parser.py	Changed parser state from single current_key to current_keys list with KEYFORMAT-aware replacement logic
openm3u8/model.py	Updated Segment class to use keys list internally with key property for backward compatibility, modified dumps() and by_key() methods
openm3u8/_m3u8_parser.c	Implemented C parser changes mirroring Python parser: current_keys list, KEYFORMAT comparison, and key replacement logic

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The by_key method performs a linear search through segments and checks key membership using in operator on a list, which is O(n*m) where n is segment count and m is keys per segment. For better performance, consider converting segment.keys to a set before the membership check, or if keys are unique objects, compare by key attributes rather than object identity.

Suggested change

	return [segment for segment in self if key in segment.keys]
	return [segment for segment in self if key in set(segment.keys)]

[bbayles]

Robot, it's strictly worse to create a new set every time.