James Tucker

https://github.com/tailscale/tailscale/pull/5915 may also have fixed a case of this for Windows users, where there's an overlap of QUIC & IPv6.

Update (2023/03/30): Chrome 111.0.5563.146 - HTTP/3 Firefox 111.0.1 - HTTP/2

Thanks for the comments @marten-seemann ! Related issues https://github.com/tailscale/tailscale/issues/311 and https://github.com/tailscale/tailscale/issues/8102. Yes, we need to implement PMTU behaviors, both on the inside layer and outer layer. On the inside we...

The "common" issuer is something specific: it enables an application to be _initially non-specific_ about the sign-in. If an application wishes to provide a global SSO login entrypoint, it can...

> If you go through the "common" or "organizations" flow, do you have any idea what the correct issuer in the ID Token would be? yes, for "Personal Accounts" it...

Attempting to reproduce, testing with a Google Fi (T-Mobile backed) connection from an iPhone to a Macbook here: ``` en0: flags=8863 mtu 1500 options=6460 inet 172.20.10.2 netmask 0xfffffff0 broadcast 172.20.10.15...

> 7\. I attempted to ping the Plex server via Terminal and observed the result of `ping: sendto: Message too long`. The command I used was `ping -D -s 1252...

I have found that it's the first window per process that has this problem, opening a second window is fine.

I guess we better move, even if this gets reverted just to avoid the trap. I believe some users have a light dependency on the number, I'd want to dig...

I work on tailscale and wireguard. Ideally what I want is a way to explicitly exclude these interfaces from consideration. We can play whack-a-mole with the plethora of userspace daemons...