On 2014-03-04 10:07, Martin Thomson wrote: > On 4 March 2014 08:57, Adam Bergkvist <adam.bergkvist@ericsson.com> wrote: >> I agree that what you describe is totally normal. The thing I was referring >> to was the conditional side effect of close(). That the channel should be >> removed from the PeerConnection's channel list if it was never accessed via >> that list, otherwise not. That may have been a misunderstanding from my side >> btw. > > I don't think that this needs to be conditional in that way. You can > break the links, or not, but there is no need to do so conditionally. > I am perfectly happy with the behaviour regarding these collections > after close() to be undefined in spec. > > On the one hand, you can just close the connection and render it > impotent. Channels already created would still be accessible, but > useless. > > On the other hand, you could also flush the collection to prevent > later access to channels (of those that the application doesn't > already have a saved reference). > > Seems like unimportant detail to me. Thanks for the clarification. Regarding the second case. I think the flushing should depend on the channels state (closed) alone, not if the app has any other references to it. /AdamReceived on Tuesday, 4 March 2014 09:19:38 UTC
This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:17:55 UTC