[gnutls-devel] GnuTLS | Minor inaccuracy in gnutls_record_send() documentation? (#806)

[Development of GNU's TLS library]

Michael Catanzaro commented:


I tried actually using this, hoping it would solve https://gitlab.gnome.org/GNOME/glib-networking/issues/15 for me, but it didn't work as expected. Problem is that some data queued from the previous gnutls_record_send() could have actually been sent already and is no longer in the send queue. So using gnutls_record_discard_queued() does not "negate" an interrupted/cancelled gnutls_record_send().

i.e. I think gnutls_record_send() actually sends data even when it returns a negative error code. As long as applications follow the old rule to always resend the same data, it was fine. But any application using gnutls_record_discard_queued() surely needs to know how much was actually sent.

I'm not sure what you would consider the intended API behavior to be. Do you think this is worth a new issue report?

-- 
Reply to this email directly or view it on GitLab: https://gitlab.com/gnutls/gnutls/issues/806#note_221366456
You're receiving this email because of your account on gitlab.com.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.gnupg.org/pipermail/gnutls-devel/attachments/20190924/acbe6476/attachment.html>