Dear Nick, thank you so much for your help. The test program is working now.
On 11/05/2012, Nick Mathewson <ni...@freehaven.net> wrote: > * In server_read_cb, you do an evbuffer_add() to queue data to be sent > followed immediately by a bufferevent_free() on the same buffer. > That's not a safe pattern: bufferevent_free() does not guarantee how > much data (if any) is sent before the bufferevent gets closed. That was my principal mistake. In particular evbuffer_add() does not per se make a callback pending. This is clear to me now, but maybe you want to insert a line like "However, existing data in buffers is not necessarily guaranteed to be written before the call to close()" before the last sentence in the description of bufferevent_free() in your book, for the sake of other newbies. > * In server_read_cb, you don't handle partial writes. There is no > guarantee that the entire message from the client, including the > newline, will arrive in a single readv call. (When it's all on > localhost, it *should* work ok with a few small messages, but it's not > a habit to get into) Well, I do handle them, by aborting the program if they occur ;) What *is* missing, however, is a protocol sync check, that no client writes several lines at once. > * You're using LEV_OPT_DEFERRED_ACCEPT, but your protocol doesn't > entitle you to do that. LEV_OPT_DEFERRED_ACCEPT tells the OS kernel > that it does not have to not treat the connection as having been > accepted until the client actually writes something. Your clients > don't write anything until they *get* a message, so Linux doesn't need > to tell Libevent "hey, you could accept some more sockets" until they > do. Yes, I completely misunderstood the purpose of that option. Thanks for the clarification. Thank you again for your help. Best regards, Alexander *********************************************************************** To unsubscribe, send an e-mail to majord...@freehaven.net with unsubscribe libevent-users in the body.
