Mercurial > pidgin
changeset 17252:a2edef5eb1b1
- Document some weird-looking logic in the GnuTLS plugin.
| author | William Ehlhardt <williamehlhardt@gmail.com> |
|---|---|
| date | Fri, 25 May 2007 05:22:25 +0000 |
| parents | 2d2c72f70e8c |
| children | c76bee45540a |
| files | libpurple/plugins/ssl/ssl-gnutls.c |
| diffstat | 1 files changed, 12 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- a/libpurple/plugins/ssl/ssl-gnutls.c Fri May 25 04:34:43 2007 +0000 +++ b/libpurple/plugins/ssl/ssl-gnutls.c Fri May 25 05:22:25 2007 +0000 @@ -123,6 +123,18 @@ gnutls_data->handshake_handler = purple_input_add(gsc->fd, PURPLE_INPUT_READ, ssl_gnutls_handshake_cb, gsc); + /* Orborde asks: Why are we configuring a callback, then + immediately calling it? + + Answer: gnutls_handshake (up in handshake_cb) needs to be called + once in order to get the ball rolling on the SSL connection. + Once it has done so, only then will the server reply, triggering + the callback. + + Since the logic driving gnutls_handshake is the same with the first + and subsequent calls, we'll just fire the callback immediately to + accomplish this. + */ ssl_gnutls_handshake_cb(gsc, gsc->fd, PURPLE_INPUT_READ); }