|
From: | Kieran Mansley |
Subject: | [lwip-devel] [task #10088] Correctly implement close() vs. shutdown() |
Date: | Tue, 16 Feb 2010 15:30:15 +0000 |
User-agent: | Opera/9.80 (X11; Linux i686; U; en) Presto/2.2.15 Version/10.00 |
Follow-up Comment #4, task #10088 (project lwip): I'm happy with changing the meaning of tcp_close() as (i) I think that the new meaning is what most callers wanted and (ii) we'll properly document the change. The only workable alternative would be to remove tcp_close() altogether and replace with two new functions with different names (but I'm not sure what) so at least anyone updating would be painfully aware that tcp_close had changed. One subtlety that seems to have got lost with closing the other bugs and opening this one is that close() should also send a RST if there are any data already received but not yet delivered to the application. To do this we need to make additions to the sockets and netconn layers to allow access to "how much data is buffered on receive for this connection" at each layer, so we can decide what the appropriate behaviour is. _______________________________________________________ Reply to this item at: <http://savannah.nongnu.org/task/?10088> _______________________________________________ Message sent via/by Savannah http://savannah.nongnu.org/
[Prev in Thread] | Current Thread | [Next in Thread] |