Bernard Normier
Bernard Normier
> If it's not used anywhere, do we give up on this PR? This PR updates ConnectionI to use it in one spot. I am in favor of merging this...
No, we should standardize on deprecated, not add more support for 'deprecate'.
> It looks like a connection that started reading a message (but didn't fully read the payload) is considered inactive. This is both accurate and the correct behavior. Merely receiving...
> We already created a new exception so I would add a new exception instead of re-use this exception which is a protocol exception. How many exceptions do you want...
> > We already created a new exception so I would add a new exception instead of re-use this exception which is a protocol exception. I added a new exception,...
> Correct me if I'm wrong but this is the case for the IceRPC ice's protocol implementation but not for the icerpc protocol implementation. In the ice protocol implementation we...
> we reschedule the timer once all these conditions are true: _writeStream.b.empty() == true, there's no more data to send _readHeader == true, there's no more message to read and...
> Nothing to suggest because I think it's a mistake to have a specific behavior for request messages. There are not many types of protocol frames with Ice: request/batch request,...
@bentoi will open a PR to improve this implementation.
## Proposal 1. Remove all Ice.ACM and ObjectAdapter.ACM properties and replace them with: ``` Ice.IdleTimeout # corresponds to IceIdleTimeout in IceRPC, with same default (60 seconds) Ice.InactivityTimeout # corresponds to...