Ian,

Thank you for the clarification. I missed that. With your proposed changes, the 
code is easier to understand (at least for me).

https://github.com/quicwg/base-drafts/pull/4409

Timo

> On 24. Nov 2020, at 19:19, Ian Swett <[email protected]> wrote:
> 
> It's a bit subtle, but the current pseudocode does the right thing, because 
> sent_bytes is only set to a non-0 value in OnPacketSent() if the packet is 
> in_flight:
> sent_packets[pn_space][packet_number].sent_bytes =
>       sent_bytes
> 
> Given the way it's used, maybe it makes sense to rename sent_bytes to 
> inflight_bytes? Though that could be confused with byte_in_flight?
> 
> Ian
> 
> On Tue, Nov 24, 2020 at 2:09 AM Timo Völker <[email protected]> 
> wrote:
> Hi all,
> 
> The pseudo code in the recovery draft uses the function 
> DetectAndRemoveAckedPackets. Is my assumption correct that this function 
> should return a list of all newly acked packets (including these that do not 
> count as in flight)?
> 
> If so, it seems there is a bug in the code, because this list is passed to 
> the congestion controller, which uses *all* the packets in the list to 
> decrease its bytes_in_flight.
> 
> Timo

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to