Blame
| 6f02e7 | Samuli Seppänen | 2025-01-29 06:40:08 | 1 | # Development Process |
| 2 | ||||
| 3 | ## Background |
|||
| 4 | - The first phase of the development process (work done in "testing") is handled properly. |
|||
| 5 | - The last phase of the development process, official releases, are handled by James, but the process/requirements for a release are not documented. |
|||
| 6 | - Right now, "testing" is effectively a fork of James' "stable" SVN tree. |
|||
| 7 | - There is no roadmap for the next 2.x release. |
|||
| 8 | ||||
| 9 | ## Which Tree to Base Releases On? |
|||
| 10 | ||||
| 11 | ### Option 1: Basing Releases on James' "Stable" Tree |
|||
| 12 | - There's no process to move code from "testing" (git) to "stable" (svn). |
|||
| 13 | - The release process is not documented. |
|||
| 14 | ||||
| 15 | ### Option 2: Basing Releases on David's "Testing" Tree |
|||
| 16 | - Currently, James makes his modifications to his SVN tree and David pulls his changes to "testing". In this regard, James' SVN tree is no different from any other external tree. |
|||
| 17 | - Only David's "allmerged" branch contains all code (James', external trees, etc.). |
|||
| 18 | - The "testing" tree should get the widest testing, especially after we start releasing testing snapshots and linking to them from openvpn.net. |
|||
| 19 | ||||
| 20 | ## How to Verify Stability of the "Testing" Code (for Moving to "Stable" or Prior to a Release?) |
|||
| 21 | - Publishing or linking to "testing" releases on openvpn.net is essential to get wider use for "testing". |
|||
| 22 | ||||
| 23 | # Other |
|||
| 24 | ||||
| 25 | ## Bridging Issues |
|||
| 26 | - From [OpenVPN FAQ on Bridging](http://openvpn.net/index.php/open-source/faq.html#bridge1): "Another bridge disadvantage should be that layer2 is insecure by design, opening your layer2 exposes to ARP poisoning and the like." |