Something is being withheld, and I'm not sure what, and that's how conspiricies grow.
They made a new release of a stable operating system. It made some nand improvements, which is sort of a vague low-level sounding tweek.
Somehow, this breaks a user manager feature, which should be way higher level feature; far opposite end of the OSI model scale actually. What connects usermanager to nand memory control is a whole bunch of operating system stuff in the middle.
Either this release is a minor improvement not worth upgrading, or there are complex changes that affect usermanager that are not documented in the changelog are too difficult to fix before releasing it. Are these changes so very important to users that it's worth breaking these usermanager features? Being able to backup strikes me as pretty important.
In addition to better automated testing, and more detailed changelogs, I would support the idea of classifying the importance of changes.
Check how HP does it. Their release notes are certainly overkill, but the describe the new enhancements, and also have a log of software fixes describing the problem and whether it can crash the switch or juts hinder a feature.
http://cdn.procurve.com/training/Manual ... 906003.pdf