Yes indeed, a change was made to limit the number of significant digits for the main stats in the report. This was done for two reasons:
Reporting 8+ significant digits is inherently misleading. While the computer is perfectly happy to calculate these digits, it gives a false sense of precision in the calculations. When we are dealing with whole read counts, often <100, calculating fold changes (especially log fold changes) out that far is not really meaningful. For confidence statistics, while we retain the magnitude in the form of "e-07" etc, the specific digits after that can't really be taken seriously. (Indeed, calculations of p-values beyond about the e-04 and lower rate should be taken with a grain of salt). Basically, too many digits encourages an illusion of precision that is not real.
When each number has 8 digits to the right of the decimal place, the vast majority of information contained in the report is actually insignificant, making it more difficult to focus on the meaningful aspect of the values. It is certainly a lot easier on the eye to look at tables of numbers with fewer digits!
I would consider adding a feature in future allowing you to specify the number of significant digits as this would be an easy feature to add.
