We've been working hard on improving the out of memory (OOM) handling by the OS since the update9 opt-in was released. A portion of our time went into demo preparations for Slush and of course the tablet launch. This has left us lagging a bit behind the internal release schedule for update10.

We did a post-mortem on late releases to figure out what went wrong with our development cycle and surely do not mind sharing our thoughts on this here. One of main reasons has been that we underestimated the work needed to successfully complete the Qt5.2 transition. We also made a conscious decision to leave for summer vacation together, leaving the development branch open for contributions without integration promotions (QA) to testing repository for that period. When we returned, it took double the time to stabilize the development branch before going forward. Coupled with losing a few of our fellow sailors, our developers had to stretch beyond their comfort zones to debug and attempt finding the root cause of the crashes, performance hits, connectivity issues, memory handling among other things. We have moved slowly, but steadily.

On a positive note, we have amended our integration process to avoid the pitfalls we faced during this transition. We have also enabled better ways to develop new features and technology in parallel with the main codebase, allowing integration only when code is relatively stable.

As for content of update10, this post is not the complete changelog. It just aims to update you on the reason we flagged update9 as suitable for opt-in.