On Week 52
skyjake:
Last week we did a very large code merge to unify our development branches into the master branch where the automated builds are made from. With over 1700 commits this was the biggest upheaval of the codebase we've seen in recent years. While this merge will be causing some side effects in the biweekly builds, it is a very important step forward on our roadmap. We are finally ready to start working in a way that allows us to make releases more rapidly and with constantly improving quality.
As we're now switching to the Candidate release phase in which 1.9.7 is prepared for stable release, I've been updating the wiki pages accordingly. The roadmap diagram has been revised along with the upcoming release notes for 1.9.7. The latter will be fleshed out gradually as we near the stable release. The Candidate phase also means that a feature freeze is in effect -- the goal is to get the stable release completed as quickly as possible; focus should be solely on fixing the existing code so that it works as intended.
My plan for this week is to verify that Snowberry works on all platforms (particularly the revised -game option and the dedicated mode) and fix multiplayer regressions caused by the code merge.
danij:
Last week was all about preparing and merging the ringzero branch into the master. As skyjake has already covered that, I'll instead focus on my plans for the forthcoming week.
Those of you who have tried the latest release have no doubt noticed that there are some regressions with addon loading. These regressions are the result of the current hybrid file loader mechanism, which bridges the old resource searching algorithm with the new resource namespace model. Addressing these regressions is top of my todo list for this coming week.
Last week we did a very large code merge to unify our development branches into the master branch where the automated builds are made from. With over 1700 commits this was the biggest upheaval of the codebase we've seen in recent years. While this merge will be causing some side effects in the biweekly builds, it is a very important step forward on our roadmap. We are finally ready to start working in a way that allows us to make releases more rapidly and with constantly improving quality.
As we're now switching to the Candidate release phase in which 1.9.7 is prepared for stable release, I've been updating the wiki pages accordingly. The roadmap diagram has been revised along with the upcoming release notes for 1.9.7. The latter will be fleshed out gradually as we near the stable release. The Candidate phase also means that a feature freeze is in effect -- the goal is to get the stable release completed as quickly as possible; focus should be solely on fixing the existing code so that it works as intended.
My plan for this week is to verify that Snowberry works on all platforms (particularly the revised -game option and the dedicated mode) and fix multiplayer regressions caused by the code merge.
danij:
Last week was all about preparing and merging the ringzero branch into the master. As skyjake has already covered that, I'll instead focus on my plans for the forthcoming week.
Those of you who have tried the latest release have no doubt noticed that there are some regressions with addon loading. These regressions are the result of the current hybrid file loader mechanism, which bridges the old resource searching algorithm with the new resource namespace model. Addressing these regressions is top of my todo list for this coming week.