On Mon, Dec 7, 2015 at 3:09 AM, Ilari Liusvaara <ilariliusva...@welho.com> wrote:
> This came up while writing serializers/deserializers for various TLS > 1.2 and 1.3 stuff... Didn't see issues/pull requests for any of > these... > > 1) ServerConfiguration has field early_data_type, which is of type > EarlyDataType. I don't see definition of that type anywhere. I > guess it was missed when that type was removed? > Oops. This got removed. Will re-add. > 2) If ServerConfiguration contains an unknown extension, how should > the client behave? Ignore the unknown extension, ignore the entiere > ServerConfiguration or send a fatal alert? > Ignore the unknown extension seems consistent. 3) Could the group and server_key fields be combined into a single > field of type KeyShareEntry? The layout would be identical, and > one could clearly reuse KeyShareEntry parsing This is a great idea. I will do it! > (especially > considering some groups have 2-byte interrior length fields > as opposed to 1-byte). > I noticed this in my implementation as well. I actually propose to just remove that extra interior length field. Can't see any reason to keep it. BTW: The new TLS 1.3 extensions don't have even test values > assigned. The values I put in were: > - early_data: 13000 (0x32C8) > - pre_shared_key: 13001 (0x32C9) > - key_share: 13002 (0x32CA) > - cookie: 13003 (0x32CB) > I suggest 40-44 in the order you list above, since we've agreed not to have specified ranges any more. -Ekr > > > -Ilari > > _______________________________________________ > TLS mailing list > TLS@ietf.org > https://www.ietf.org/mailman/listinfo/tls >
_______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
