I fixed the following bugs:
#2789 multivariate polynomials over residue fields of number fields are broken
#1367 weird bug creating fractional ideal in relative number field
That "weird bug" was actually lots of missing conversions and questionable decisions over the years. This required a great deal of time to understand the infrastructure to fix properly, but opens the door for easy slaying of additional bugs, including...
#4869 make element of relative number field from polynomial
#4727 list method on relative number field elements is broken -- it doesn't satisfy the most basic consistency check
I reviewed several patches at the beginning of the conference, too. Most made it into 3.3.0alpha2.