On Mon, 15 Jun 2009, Phillip Wu wrote: > Is it not possible for a non-blocking read to return 0 when there is no data > to read but in actual fact the remote side has sent an EOF?
Oh, right you are! > In non-blocking mode when I get a 0, I could confirm that eof has been sent > by calling libssh2_channel_eof. There could be a slight chance that I could > loose data during the time delay from the eof call. How would you lose data because of this? > So I think it would be slicker to return LIBSSH2_ERROR_CHANNEL_CLOSED rather > than 0. > > What are your thoughts on this? I agree with you now. I think. But we must make sure that we properly drain the incoming data "queue" first before returning LIBSSH2_ERROR_CHANNEL_CLOSED. As you're onto this, will you write a patch and see that it works for your use-case first? -- / daniel.haxx.se ------------------------------------------------------------------------------ Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects _______________________________________________ libssh2-devel mailing list libssh2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libssh2-devel
