Repository navigation
Conversation
BaseKeyProvider always derived AES-128 frame keys. The prefixed WebRTC Android library already accepts the key size (a trailing int, in bits, defaulting to 128 in the shorter overload of createFrameCryptorKeyProvider), so this only forwards it. Adds KeyProviderKeySize (BITS_128 default, BITS_256) as the last BaseKeyProvider constructor parameter, so existing callers are unchanged. Every participant in a room must use the same size. Matches KeyProviderOptions.keySize in the JS SDK and the Swift SDK (#1200). Fixes livekit#1060 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011bMyZt4jGEfw4YQ5siiMRs
|
|
🦋 Changeset detectedLatest commit: efbc333 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
Adds a short example for the new BaseKeyProvider keySize option to the e2eeOptions documentation, as the repository guidelines ask for new public APIs. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011bMyZt4jGEfw4YQ5siiMRs
|
Thanks. We kept the new parameter last with a default so source compatibility is unchanged. BaseKeyProvider's constructor has grown the same way before (e.g. keyDerivationAlgorithm in #886), so we followed that pattern rather than adding hidden overloads. Happy to add a hidden deprecated constructor with the previous signature if the maintainers prefer binary compatibility here. |
|
Thanks for working on this. I'm already working on an app which uses E2EE and would like this very much ported in android (Web & Swift already provides it, swift-recently merged). Would very much like this feature enabled. Happy to test it out when available. |
|
LGTM, @netbrut can you sign the CLA? |
|
@netbrut I'd rather not let this one go just over the CLA bot not working for you, can you just comment here and say you've signed the CLA? I think that should be fine enough for our policies. |
|
Sorry for closing this, my mistake. I have signed the CLA. The CLA bot stays pending because the commits were authored under the git identity "claude", which maps to an unrelated GitHub account, so it can't be signed for. I'm the author of this contribution and I agree to the CLA for it. Happy to re-author the commits under my own account if you'd prefer the check to go green. |
|
I've confirmed with our team that just commenting here is good enough to cover it, so you're all set. Thanks for the PR! |
Fixes #1060. Android counterpart of the Swift change in #1200.
BaseKeyProvider always derived AES-128 frame keys. The prefixed WebRTC library already used by this SDK (150.7871.04) accepts the key size: FrameCryptorFactory.createFrameCryptorKeyProvider has an overload with a trailing int (bits), and the shorter overload passes 128. This PR only forwards it.
• New KeyProviderKeySize enum (BITS_128 default, BITS_256), added as the last BaseKeyProvider constructor parameter, so existing callers are unchanged.
• Default key provider in E2EEOptions explicitly uses 128.
• Changeset (minor).
Every participant in a room must use the same size; the KDoc says so.
Verified: :livekit-android-sdk:compileDebugKotlin passes. Not verified: runtime behavior (needs a device) and JS ↔ Swift ↔ Android interop with keySize: 256. I'm happy to run the interop test as soon as a release is out.