Tomasz Sterna wrote: > Dnia 03-10-2007, Śr o godzinie 10:30 -0600, Peter Saint-Andre pisze: >> Would anything break if we specified that the receiving entity SHOULD >> close the stream and then the initiating entity shall send a new >> initial stream header? > > I do not know any implementation that handles this closing tag today. > The implementations I know would just process the closing tag through > TLS layer and fail miserably. > So things will break. > > >>From my standpoint we really do not close the stream. The previous state > (auth, compression, encryption) is preserved and we just add another > layer to the stack. > It's more like opening another stream inside a context of previous > stream. :-)
Just to be clear, I don't want to change this behavior. Someone poked me about this in the jdev room and I promised that I would bring it to the attention of the list. Given that no implementations behave this way today, I'd agree with Dave Cridland that we should change "it is not necessary to send the closing stream tag" to "an implementation MUST NOT send the closing stream tag". Peter -- Peter Saint-Andre https://stpeter.im/
smime.p7s
Description: S/MIME Cryptographic Signature
