Basic info

Topics

Updated: 2.7.x release
Version 2.7.9 is expected to be released tomorrow 4 November.
It will again have a number of security fixes in there.
4 weeks following that yet another release will be done.
New: add openvpn-rfc repo to gerrit for reviewing
Currently changes to the OpenVPN RFC documents are reviewing on github.
There is a strong preference to do this on gerrit like other openvpn2 patches.
uddr35 said he will look into it.

Updated: 2.7.x release
Originally 2.7.8 was planned for last week, but a last moment fix for Windows DCO driver made us decide to postpone till today.
2.7.8 with the Windows DCO driver update will go out today.
Version 2.7.9 is now planned for 4 November.

test latency and MTU in openvpn
A new feature is proposed for adding to master; ability to do latency and MTU tests in the OpenVPN protocol.
This would allow for a client to for example select the best available server based on test results.
This is currently under review: https://gerrit.openvpn.net/q/topic:%22oob-server-probe%22
The design document is here: https://github.com/lstipakov/openvpn/blob/server-probe-design/design-server-probe.md
2.7 security audit
The security audit was completed for 2.7.0 and results are available internally.
We are working to publish this in an official blog post. ordex and novaflash are driving this in OpenVPN Inc. internally.

Coordinating big changes
A related discussion is how to handle big changesets (e.g. currently the multipeer changes from plaisthos and the oob changes from lev). There was some unhappiness that these changesets are dropped without sufficient context and that makes it hard to keep track of all the developments. One advise is that developers that want to have big (or many) changes merged are encouraged to write a mail to the mailing list to provide a "cover letter" for the patch set and a convenient place to discuss the overall direction without getting bogged down in the discussion of individual changes. We have lost this practice since not generally using mail-based patchsets anymore. That can be also a good opportunity to point out dicussions happening in PRs for the RFC document.