Commit e06289

2025-09-03 12:55:33 novaflash: -/-
/dev/null .. Meetings/2025/2025-09-03.md
@@ 0,0 1,98 @@
+ # Basic info
+
+ * Time: Wednesday 3 September 2025 at 14:00 CEST (12:00 UTC)
+ * Place: #openvpn-meeting channel on !LiberaChat IRC network
+
+ # Topics
+
+ - **Updated: OpenVPN community meetup 2025**
+ https://community.openvpn.net/openvpn/wiki/CommunityMeetup2025
+ When: 25 and 26 october (Saturday and Sunday), with travel days on the Friday before and Monday after.
+ Where: Naples, Italy.
+ Meeting room: https://www.hotelparadisonapoli.it/en/home-page.aspx
+ Hotel: Paradiso Napoli Hotel.
+ Beer: yes.
+ T-shirts: yes.
+ Invites were sent to openvpn-devel and some selected individuals. If we missed anyone please reach out.
+
+ - **Updated: forums situation**
+ minx offered to set up a new forum and help us migrate old data to it.
+ We have a new setup with flarum at a hidden location and we're trying to get the old forums data in there.
+ The tool for that called nitroporter had some difficulty, partially because the database is just so huge, but there is some progress.
+ Sad news; minx left openvpn inc. however, he does still want to try to get this working and is helping us still.
+ Once the migration is done, we can run some tests, and see if it meets our standards. Then we can plan a cutover date.
+ To the question where the instance should be hosted, community indicates it should be under the community AWS infrastructure.
+ To the question what authentication system should be used, community indicates it should just be the built-in system from the forum solution itself.
+
+ - **Updated: wolfSSL license changed from gplv2 to gplv3**
+ The release notes of wolfSSL indicate the license changed to GPLv3. This basically makes wolfSSL incompatible in regards to licensing with OpenVPN.
+ An issue was opened on github and we received a response to give them some time to come up with a solution https://github.com/wolfSSL/wolfssl/issues/9143
+
+ - **Release 2.7**
+ DCO for Windows multipeer statistics driver part is done - userspace part is waiting review.
+ The FreeBSD patch for float was merged, should show up in 14.4 probably.
+ Planned release date for 2.7 beta1 updated to September 4th.
+ These are some known remaining tasks:
+ lev is working on broken "bytecount" functionality when using Windows DCO in server mode.
+ Epoch data keys for Linux DCO and Windows DCO can be done in a minor release after 2.7.0.
+ However ralf_lici is already looking into adding it into the Linux DCO right now.
+ PUSH_UPDATE review and server-side support - cron2 wants to discuss changes.
+ mi prefix handling, int/uint fixes, and the new 'real route' gateway handling.
+
+ - **t_server and t_client testing framework**
+ mattock reports progress on building automated testing environment for really old OpenVPN versions.
+ This covers tests for versions going back to 2.2.
+
+ - **push_update / live route updates**
+ Client-side support for push_update is now merged.
+ For server-side support, company did QA on it with openvpn2 but found some missing features that are being added now.
+ Client-side support made it into alpha3.
+
+ - **format code using clang formatting**
+ It seems prudent to do this before the real 2.7 release.
+ It was discussed and there's some additional work to be paid either on reviewer or submitter side.
+ Strategy will be; get all code into beta1, collect bugfixes, reformat everything, then beta2.
+ Collect more bugfixes and then do release 2.7.0 and branch it off into its own release branch.
+
+ - **cve.mitre.org deprecation notice**
+ The MITRE CVE site we are linking to for all CVE record references has a deprecation notice. Instead cve.org is being advised as the site to use from now on.
+ novaflash made an internal company ticket to update links on the main site.
+ novaflash search/replaced all the CVE links on community wiki to the new address.
+
+ ## Backlog
+
+ - **build failures**
+ Changes in Gerrit do not automatically pick up fixes from master, so sometimes we see a lot of build failures even though those are fixed in master.
+ Some options were discussed, seems like the one preferred is to reject pushes that are behind. djpig will look into it.
+
+ - **Reference manuals OpenVPN 2.x**
+ There was a mistake made on the main website which for a few hours? or days? made the reference manual 404.
+ This has been restored. However the URL will slightly change in the future, but a redirect on the old URL will be in place.
+ mattock raised the question where the latest OpenVPN man page should be hosted.
+ When discussing this we agreed that djpig will contact the technical writer for openvpn.net and see if the ingestion/updating process can be smoother.
+ Independently we can also investigate getting the manuals on the wiki somewhere, perhaps automated.
+
+ - **snapshot releases via Chocolatey software**
+ mattock contacted the Chocolate package maintainer for OpenVPN and asked if he would be okay with publishing Windows MSI snapshots as well.
+ Seems like the maintainer is amenable to helping us achieve that goal.
+
+ - **2.7 security audit**
+ ordex mentioned that OTF offers the possibility to get a 3rd-party security audit for supported projects. So we will apply for that around or after the 2.7 release to review the latest code.
+
+ - **Tunnelcrack progress (see TunnelCrack community wiki article)**
+ Status update on TunnelCrack mitigations:
+ The tunnelcrack mitigation for Windows has gone in master, which will go to 2.7 release. There is the possibility for it to go to 2.6.x if we can find testers for this.
+ Windows, openvpn2: block-local merged to master, not to 2.6.x. openvpn3: in code review.
+ Linux, openvpn2: in progress. openvpn3: in progress.
+ macOS: to be determined.
+ iOS: to be determined.
+ Android: not vulnerable.
+
+ - **donation collection**
+ From earlier exploration it is clear that setting up a legal entity is not worth the expense at this point. We're just starting out with donations.
+ What we can do is start out with an existing company that can collect the money and puts it to good community use. ordex volunteers to take this on.
+ There are some options to consider. There may be existing solutions that we want to consider.
+ PayPal seems overly expensive with all their fees.
+ Stripe could be worth considering for credit card processing.
+ GitHub Sponsors was mentioned as a possible solution, this is worth investigating.
+ Open Collective was also mentioned, that needs some investigating how that exactly would work for us.
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9