libevent / libevent-book Goto Github PK

at line 241 of Ref2, the flag for O1 event notification shall be EV_FEATURE_O1 instead of EV_FEATURE_ET ?

if ((f & EV_FEATURE_ET))
    printf(" Edge-triggered events are supported.");
if ((f & EV_FEATURE_ET)) <- EV_FEATURE_O1 ?
    printf(" O(1) event notification is supported.");
if ((f & EV_FEATURE_FDS))

it is useful and I wanna take a try. Is it OK?

In

https://github.com/nmathewson/libevent-book/blob/ab4dd1107bbfe419e94ccabbe9473b542173bccd/01_intro.txt#L5

why named asynchronous IO?

In my opinion, asynchronous IO is use like aio_write, aio_read , but libevent use like epoll , so there should use synchronous nonblocking IO

Notice a broken link on http://www.wangafu.net/~nickm/libevent-book/TOC.html

• link:Ref8_listener.html[R8: Connection listeners: accepting TCP connections]
• link:Ref8_dns.html[R9: DNS for Libevent]
  ^^^^^^^^^^^
  Shall be Ref_9__dns.html ?

The "Working with events" sections mentions user-triggered events, but then doesn't discuss a canonical way of having user-triggered events.

Can this be added?

Here: https://github.com/nmathewson/libevent-book/blob/master/Ref6_bufferevent.txt#L271

The sockaddr_in is memset and the sin_addr and sin_port are set, but sin_family is not, leading to error 47 (Address family not supported by protocol family).

Hi,

https://github.com/nmathewson/libevent-book/blob/6eb2799fdff4d53235d21dccca47ffc66513d250/Ref5_evutil.txt#L339-L340

I think that the following is right.

The return value from evutil_inet_pton() is 1 on success, -1 on invalid af, and 0 on invalid address format.

Hey,

This bufferevent_base_connect() function is mentionned in the book around here https://github.com/nmathewson/libevent-book/blob/9e9e5242f6fcd4cc6bfe564c22756ab6f5ce5914/Ref6_bufferevent.txt#L296
I can not find any other occurence of it in the book or in the libevent api.
Should I be reading bufferevent_socket_connect() instead ?

Seems like there is a problem with asciidoc 8.6.3

  $ make pdf
  a2x -f pdf LibeventBook.txt
  a2x: ERROR: xmllint --nonet --noout --valid /Users/tcurdt/Desktop/libevent-book/LibeventBook.xml returned non-zero exit status 4
  make: *** [LibeventBook.pdf] Error 1

In nmathewson/libevent-book/blob/master/Ref4_event.txt under _constructing_signal_events there is a warning about setting timers from signal callback "It might not be supported".

I did a SIMPLE test and did not run into a problem, furthermore it would make sense for this to potentially be a problem in the signal handler itself, rather than the callback. I assume that the callback should be executing in the event-loop context and thus should be okay. In either case, IMO this should either go away or become more definitive, i.e. list cases, environments where it is not supported.

Thanks
-Nevo

In file "01intro.txt"

/* Add all of the interesting fds to readset */
for (i=0; i < n_sockets; ++i) {
if (fd[i]>maxfd) maxfd = fd[i];
FD_SET(i, &readset);
}

Should it be FD_SET(fd[i], &readset)? I'm not sure about it.